As mentioned in my last entry I don't know much about IPMP (Intelligent Platform Management Interface), so I've done some research. I have seen the accronym before but I guess I just ignored it. IPMP looks like an abstraction layer between the operating system and the motherboards (Intel says "baseboard") monitoring systems currently accessed directly via i2c or SMbus. Anyone who has run Linux and wanted to get CPU temps knows alittle about this via lm_sensors, where you access both i2c and SMbus buses in order to obtain information, which however is controlled by drivers and configs on a per chipset basis. The old problem with lm_sensors was that if they didn't have a driver for your manament chip (usual a Winbond) then you were out of luck untill they did (or you learned to write one yourself). So it seems that with IPMP you have a hardware provided abstraction that lets you send IPMP commands to the IPMP management chip and it all just gets sorted out properly. To some degree it looks like the system monitoring equivelent to SNMP, except instead of sending requests to another servers SNMP agent you send a command to the IPMP processor on your motherboard. There is already a OpenIPMI implementation which currently runs on Linux but was written to be as portable to other OS's as possible.

Anyway, it looks like the X86 based LOM's on the new Sun boxes are based on IPMI which makes the addition to Solaris a huge huge addition, this means that Sun Control Station (which I only just now learned of) should work with Solaris fully. If the API is nice and clean this should enable the creation of some simple and powerful system monitoring applications to really give us some extra power over our systems. I'm looking forward to seeing the interface in Solaris10 (hopefully tonight!).

To learn more about IPMI yourself, check out these links:

• Intel's IPMI Page: Intel is the creator of the IPMI standard, so this is the #1 page to check out.
• The OpenIPMI Project
• The GNU FreeIPMI Project
• The IPMIutil Project (a componant of, don't laugh, Carrier Grade Linux... ok, laugh now.)
• Learn about the IPMI capabilties of the v20z
• i2c Refresher
• Some good articles on SMbus and i2c
• A list of IPMI commands avalible on the v20z and v40z LOMs
• LOM interaction from Sun Control Station discussed here

w00t! Solaris Express 2/05 is up and ready to rock! It doesn't look like the Whats New or Release Notes documentation has been made avalible on docs.sun.com, or at least I can't find it, but word was (is Alan the best or what?) that iSCSI was being added, X.Org Release 6.8.2, Chelsio 10G Ethernet Driver, and more. I've been looking forward to this release.

UPDATE: Damnit...... so apparently now "Whats New" information is at the bottom of the Express page. Didn't even notice the link in front of my eyes, the worst place to put something. :) Anyway, according to the page it says Solaris x86 Kernel Selection has been added, which would be seriously kick ass. Webmin and Apachev2 were also apparently added, but Apache2 has been in Solaris10 for a while now I thought. Webmin is nice for some, but for me its just something else to uninstall. And this "Intelligent Platform Management Interface", I've never heard of it... at least I don't think I have. Anyway, hopefully I can upgrade tonight and give a review. But really, I just want the iSCSI initiator.

Two years ago it was clear, Brocade was the undisputed heavy wieght champ of Fibre Channel Fabric Switching (thats Storage Area Networks for you non-storage people). Now, I dunno, you almost get the feeling that McData is leaving them in the dust, and honestly I've been confused about who's got the upper hand. I always love an underdog but honestly I'm not sure which one is, its a tight race, and both have partnerships galore, although when I talk to storage vendors more and more I'm hearing McData being recommended over Brocade, a sign of good solid dealmaking.

So now we've got some numbers. McData chimes in with a rough year, reporting FY05Q1 revinue of $105.8 million and a net income of $6.8 million. Brocade is looking pretty, reporting FY05Q1 revinue of $161.6 million and a net income of $27.9 million. Broace's revinue was up 11% year-over-year, versus McData being down 7% year-over-year, despite being up 7% percent sequential. However, even though McData's numbers are mixed they've increased their net income 48% from the previous quarter! So, numbers are mixed in my mind. Brocade is the leader right now based on financials, but I'd like to see volume numbers and analyst market share estimates. McData's aquisition of Nishan Systems awhile back gives them an edge as iSCSI continues to come into focus, and Brocade is looking in several directions. Some tough choices loom down the road as Infiniband looks more attractive for storage, something both Brocade and McData are ready for, but we'll see how they perform when the time comes. Of all the competition between vendors out there in diffrent niche-markets I find the FC switch market to be the closest, especially because there are so few players and I don't see people flocking onto the alternative from QLogic (except Sun's OEM'ing the SANBlade line). Most storage vendors seem to be hedging their bets anyway by pushing one switch but also being prepared to sell the competition. We'll see where this goes over time.

Jim posted an interesting blog entry reguarding blogging in an article entitled "Are blogs good for business?" (no link, reg needed, see the exceprts in Jims link above). The blog thing has come up time and time again. I found the opinion of American Library Association president Michael Gorman to be an interesting one. But, I'm definately a reluctant blogger and so I just have a couple of thoughts....

I've ment to post about my thoughts on "tinkering" and just haven't gotten around to it. "Tinkering" is what I call the act of working on something with no tangable productive outcome beyond yourself. So, if you learn the ins-and-outs of WindowsNT, for instance, thats tinkering. If you learn how to assemble a Sun Cluster, thats tinkering. If you learn to write C or Ruby or some other language, thats tinkering. Tinkering falls into 2 distinct categories: waste and education. If you can personally apply your tinkering into a career path, then its productive (and how most of us got to where we are!). However, if you don't parlay that information into some personal goal then you just wasted your time. Is spending hours tuning and primping your home entertainment system useful tinkering? Not really. Is building the sweetest most bad-ass car on the road useful? Nope. Sure it brings personal satisfaction, but is that enough? In the open source world (the UNIX philosophy), no, its not. In your average day stop and think about what your doing. Ask yourself, would anything that I'm doing possibly be usefully added to the sum of human knowladge? And what is the Internet but a place to amass the sum of human knowlage? So, when you tinker with something, how can you turn that into a useful and productive event? Share that knowladge. If it takes you an hour to figure something out, why should someone else need to spend an hour of their lives figuring it out too? What if they could read a document or blog entry that your wrote after learning a subject and then save themselves the time? Wouldn't you like to learn how to do something nifty in 5 minutes instead of 5 hours? I know I would. That was the thought behind my Cuddletech Veritas KickStart tutorials. The manuals for VxVM were so poorly written and confusing that it took me literally months of pooring over them again and again and again because I didn't have the money or oppertunity to have a testing enviroment to figure it out myself, so I had to work it out in my head, and when the lightbulb finally went on I had one of those "WOW! This all makes selse!! Its so much easier than this manual lays it out!!! Damnit, why was this so hard to figure out!!!" moments. So I took that study time, that otherwise would be wasteful and useful only for my personal growth, and wrote my own documentation to get people started, saving them the time and effort that should be spent on something more productive. And its been hugely successful. Everyone is learning and growing thanks in part to my small contribution to the sum of human knowladge.

The nice thing about blogs is that where previously I would write full length papers or books on a topic I can now simply plop out into blog entries with little hassle. Besides, after hacking on something its nice to share some of my experience. Will some people mock you? Sure, but who cares, for every person that says "Pffft, thats simple stuff, everyone knows that you l00ser" there are 10 people saying "Wow! Thats kool and helpful, I'll try it!".

And you don't even need to add information into a blog that solves a problem. Take my entry on Asterisk, I just thought it was sorta nifty and mentioned it. Now I'm doing some kool things with it and looking at ways to help the project, and several other people have responded and told me that they were really glad to hear of other Solaris users working with it and even more who hadn't heard of it and are now excited to play with it. One of the most powerful things anyone can do is to let other people know that they aren't so alone as they think. While we all search to be unique individuals unlike anyone else, we also yearn to know that we're not so wierd as those around us make us feel and that there are others out there like us. Just serving as a cheerleader, if you will, can be a hugely useful contribution, encouraging others, providing useful and positive feedback to a project, and making folks who work hard at something feel appreciated. Because, at the end of the day more than money or personal gain what we all really want, especially in the UNIX and open source worlds, is to feel appreicated. As sappy as I know that sounds, its the truth. Blogs are a newish and accepted way to do this effectively and quickly without all the old school troubles of having to start a site and register a domain and all that stuff like I did with cuddletech. This is just another outlet by which to pitch in and make the world just alittle bit better for all of us... one person and a time.

If you've been on the net less than 8 years just skip this entry.

Did you know LambdaMOO is still avalible? I was shocked. Dropped the name into Google after someone brought up MUDs in IRC and was shocked to find that its still accessable! Access it via telnet at lambda.moo.mud.org port 8888. Its just as I remember it. I used to be into it way back when, but never became a heavy MOO'er, and never really got into MUDs or MUSHing. I sort of got out of things when the Furry invasion happened and the whole world of text-based interactive enviroments got popular in the pre- and early web days. Viva Gopher. Viva LambdaMOO.

My good buddy and fellow Enlightenment developer Rephorm was telling me that he's now working on some VoIP stuff using Asterisk, an open source VoIP IP/SIP PBX. So now I'm distracted because I'm reading up on and playing with VoIP. :)

If you haven't checked in with VoIP technology beyond the residential/consumer stuff (Vonage) in a while you'll be suprised at how far along its come. Asterisk is pretty sweet and massively featured at no cost. Digium, among others, provides PCI cards to connect existing T1s, PRI, POTS lines, etc, to your Asterisk PBX. VoIP's 2 popular are H.323 and SIP, the latter is becoming the favorite. Plenty of SIP clients are avalible as both "hardphones" (your desk phone) and "softphones" (software SIP client using your sound card, etc). Its all interesting enough that I'm thinking perhaps that I should write an EFL based SIP softphone, because the existing softphones for UNIX are ugllllllllly. So between inexpensive (<$60) SIP hardphones and the widespread avaliblity of ATA's (Analog Telephone Adapter, connecting RJ45 Ethernet to the RJ11 your existing phone uses, which means that ATA is also doing Analog-to-Digital conversion for you) we're definately getting somewhere. Asterisk can handle call parking (put people on hold and transfer them), voicemail, extensions, voice call trees ("For Sales Press 1 now, For PR Press 2.."), confrence calling, and a whole lot more. And its pretty simple. And, if your using Gentoo, you can just emerge it and go! And best of all, Asterisk runs on Solaris too! (I'm thinking that Asterisk under Sun Cluster 3.1 would be sweet!)

Anyway, I've got to spend some time learning the avalible SIP API's out there with hopes to write my own softphone and when tinker more with Asterisk. I think tonight I'll probly move Asterisk off my Linux system onto a Solaris Zone and go from there.

If your interested in VoIP check out these links:

• The VoIP Wiki: Everything you want to know! Start here.
• Getting Started with Asterisk
• Xten Sofphone: Freeware SIP softphone for Win32 and OSX thats pretty!
• VoIP entry on Wikipedia
• Asterisk Homepage: Open Source VoIP IP/SIP PBX
• Digium: Solutions based on Asterisk, Support, and Hardware
• "It's Time To Take a Look At SIP": Good artical on SIP
• FreeWorldDialup: Free VoIP alternative to Vonage, etc.
• SIP Implementations

Also, there is a VoIP confrence coming up in 2 weeks here in San Jose: VON2005 (VON for Voice on Net, the term used prior to VoIP). I'd love to go to that but with expo only passes running $150 I probly will have to just wait for show reviews.

Today I called Tadpole looking for pricing on their SPARC based laptops. I've been thinking about getting a lappy and with OpenSolaris I'll be wanting to do alot of demos in the future, and I personally prefer SPARC over X86, so a nice fast Viper or even better yet Bullfrog would make me oh so happy to show off OpenSolaris and Enlightenment on the road. Costs have been dropping so I was hoping a Tadpole was more accessable... but not today. I called into Tadpole Sales, got Dennis Vines, who apparently had just woken up or something. I asked for a price list, but they only go out to VARs, which means that I have to ask for pricing one by one, and he wasn't terribly forthcoming with price breakdown of the lappies so that I could find the best fit. He sort of drowned on like he was wasting his time... so apparently Tadpole's not terribly interested in doing business. This is really sad, since I've been trying to talk my employeer and some other companies into picking up some Tadpole's for demoing and as replacements for developers, since the ability to develop and demo on SPARC is really nice. So much for that. I guess I'm going to have to suck it up and just get an AMD64 lappy... too bad there aren't many with 17" screens. I asked Dennis whether Tadpole was interested in OpenSolaris or whether they'd support it in the future... apparently he'd never heard of it or just didn't want to comment, he just sighed and said that Tadpole supports whatever Sun supports, which I guess means this isn't going to provide for cross-breading between Tadpole and the OpenSolaris community to develop more robust drivers or fill in gaps for either of us. Oh well, so much for Tadpole. Hopefully their gov't sales can keep them going, since I got the feeling that its gov't sales they really want, not commercial, and certainly not personal.

Timothy Prickett Morgan at itjungle.com has put up a piece about Red Hat's recent moves entitled: "Red Hat Bashes Sun As It Launches Enterprise Linux 4". The title sorta sums it up. Sun and Red Hat have virtually locked horns by most peoples estimates, but Sun's also fighting the good fight to win HP's beatup customers, IBM refugees, and keeping up the conversation with Linux vendors, namely Red Hat. Jonathan recently said: "The number-one competitor for Solaris is not Red Hat, it is Windows. We see that market now opening up." This is an interesting return to reality, Win32 should still be the focus on our conversions trying to first and foremost save customers from the terrible, insecure, unscalable, and buggy pile of crap that is Windows. Linux and Solaris are both better options for these people, and with the help of companies like Scalix (that goldfish picture rules) who offers the best Exchange drop in replacment to date, we're getting there. Looking at other vendors who are showing Win32 customers the light, like Solinus's MailFoundry thats bringing Solaris stability and power into the server rooms of Win32 shops, we're making progress in this fight. Truly our partners are a much stronger asset than ever before. I'd rather see Win32 shops stuck in the past move to either Linux or Solaris, and then they come over you'll clearly find with some homework that Sun solutions are the best and most cost effective platforms to build on, reguardless of whether you run RHEL or Solaris10. Its choice, yet again, that we're after. Certainly Solaris is a much more robust, tested, proven, and diehard OS than Linux is, and for most business applications you just can't deny that. Definately Linux has its place, but hosting my billing database just isn't one of them. Old nitches that suited Linux well, such as webserving or almost any application where reliablity wasn't a #1 concern (ie: anything with a load balancer in front or what we now call a *sigh* "grid"), are now being slightly erroded by the lockin forced upon you by support agreements with Red Hat or Novell, and looking at Solaris10 technologies such as Zones, DTrace, SMF, and others you are given more power than ever right there on the most trusted platform around. Sun powered the internet in the beginning, and they set the standard by which (knowlingly or not) Linux has tried to live up to and attempt to eclipse, but not today. And reguardless of what Red Hat thinks, adding Solaris support to the "Red Hat Network" isn't going to steal customers, your just going prove to the world that Solaris really is the high avaliblity and robust platform that people need in their datacenter and you want to leverage your way into that base, but I don't know of many folks who are going to wanna deploy Red Hat Network just to help manage their existing enviroments. It didn't work for IBM with Tivoli, it didn't work so well for HP with OpenView and I doubt this lack-luster effort is going to do anything but point at Red Hat and say "See? Now they understand that they can't live without Solaris". If I were Red Hat I'd spend more time keeping my eye on Gentoo and hope that someone doesn't see an oppertunity to compete with 'em.

UPDATE: James Dickens is the man! He's copied the comparison table listed in the above article and added a column for Solaris10! Check it out here!.

Last night I finally got my paws on GT4 and got to put about 2 hours into it. Being a parent, the first analysis is that Nova didn't care much for the game except to sit in my lap and mash the analog sticks while I was driving rally... which she quickly got bored with, tossed the controller and brought back a book for us to read. So, don't expect GT4 to entertain your toddler. As for actual game play, the arcade mode is massively improved including the ability to "quick tune" your car prior to a race but without all the complexity in a full custom tuning session in Gran Turismo mode, you just dial up the HP and drop the weight. Also, in Arcade mode you can finally use Racing Super-Soft tires(RSS)! Driving arcade mode on medium compound tires sucked. The car list is large, 600 in all, but honestly its not so big as it seems due to the huge number of duplicates with diffrent specs (almost 100 Skylines, from antique to racing class). Still no Ferrari's, Masseratti's, Lambo's, or Porsche's (but the RUF'ified Porsches are), but then they aren't going for that crowd. Being a huge fan of European cars I immediately pass over the possibly superior Asian cars that can really help you win races.

In Gran Turismo mode the general interface has been improved. Snow modes now have been added, and an impressive number of tracks, real and imaginary. This game is worth having just to drive the amazing Nurbergring (the old 14 mile track) at blistering speeds. For GT3 players you get an edge by being able to bring over your A and B licenses as well as up to $100,000 to suppliment your starting $10,000. Using that cash I bought a Volvo S60 T5, added some upgrades and start working through licenses. Drivability is improved, meaning that its much more realistic, however the more realistic the driving gets the more you realize you really don't want to drive with a controller... making me want to pick up a GT4 wheel. Being a student of racing, kart driver, and gamer I'm finding that GT4 really makes you clash, because you can see that its trying to enforce proper driving style (forget about missing your apex and staying on track) but its just tough with a game pad. I compared notes with my brother-in-law and found that using the analog sticks for both steering and throttle/brake was a major improvement in terms of control, but it takes some getting used to. What I did find is that automatics in GT4 are useless, I switched over to manual transmition and had massively improved results right away. All in all, I'm pretty happy. I would have liked some more European cars, but honestly I really only care about my baby (the Opel Astra Team Pheonix) being avalible which it is. I'm not thinking that I need to buy a wheel from some place that will let me return it a week later, which isn't my style, I never buy something with the intent to return it... but in this case I just have got to give the wheel a try. If you like racing games, this is the best one out there. The graphics are good, but honestly too much pixelization for my tastes, giving you a headache, but maybe its not so bad in 1020i.

CBS Marketwatch is reporting on Gartner's report. The skinny: HP is #1 in shipments, Dell is #2, so on in the general market. In UNIX server sales, Sun takes the top slot as #1 with revinue of $5.1 billion, putting HP in second place with $4.9 billion (which I find slightly shocking). IBM pulls in at #3 with $4.3 billion, but we've gotta watch 'em because they topped HP in the 4th quarter and while Sun and HP share is both down (Sun by 5%, HP by 8%) IBM was up by 7% for the year. So... things are in motion. Undoubtably IBM's going to overtake HP in the next fiscal year, but what will the fate of the sector be and what will the future of Sun and IBM turn out to be? Remember folks, this is in our hands.

All questions that I've asked about Sun's future with iSCSI have gone unanswered.... but hey, not a problem now. I plopped "iscsi" into search on docs.sun.com and up came this: Chapter 15 Setting Up an iSCSI-Based Storage Network (Tasks). So, now we've got some ideas. Nothing mentioned about a target implementation, so you'll need to continue to rely on the existing Linux targets (read here). This will be a nice and welcome addition to relieve us of our dependance on the Cisco SN5400 Solaris driver that we've used thus far (version 3.3.6 still doesn't have Solaris10 support, and my attempts to get it working on Solaris10 have failed).

Coupling Zones and iSCSI will be a useful pairing in the future. Imagine being an ISP who wants to sell premium accounts to users. Instead of a login and homedir on an existing server you now offer them a zone and give them root to it. Want to reboot your virtual server? Fine, reboot your brains out. Then couple this with an iSCSI mount on a large shared storage system (NetApp FAS900, or a robust Hitatchi Thunder or inexpensive Apple XRAID attached via an McData Eclipse SAN Router, Brocade Silkworm Multiprotocal Router, or Cisco SN5400 Storage Router). In this setup you can offer end-users an enviroment of their own, under their complete control, to be used as they like. They buy an account and you give them a zone. They ask for 10GB of storage, you mount an iSCSI LUN for them and let them use any filesystem they like. With this type of setup you could host 100 or more high ASP customers using little more than a quad CPU UltraSPARCIII based system, an Apple XRAID for some huge and cheap storage, coupled with a storage router. For less than $25,000 you could charge as much as $200 or more per month, and give them a first class experience, hosting as many as 800 users accounts with this setup in a single rack... all of them thinking they have their own dedicated server with root, reboot privelages, and the works.

Side Note!!!: Infiniband junkies will wanna check out Chapter 9: Using InfiniBand Devices (Overview/Tasks).

Raster got slashdotted today. Finally some of the progress made with DR17 is being appreicated. On Raster's site (when its not being /.'ed) you can download videos of DR17 in action. You can learn more about E and its backend, the EFL, on enlightenment.org. I've gotta spend some time updating the site in the near future. If your interested in the EFL's emense power, you can check out the docs Dan Sinclair and I hacked up awhile back, including "Building Interfaces with Edje: The Edje Developers Guide" by me, "The EWL Book" by Dan, and "The EFL Cookbook" by several of us, primarily Atmos, Dan, and myself. The Edje book really needs some big time updating, but I haven't had time to work on it.

I've been scraping to find a way to play GT4 online... only LAN play is supported. I've figured that I'd just setup a VPN and not worry about it, but I found this: XLink Kai. Basically it handles the VPN'age for you and lets you play amoung a large group of users. Quick Start guide is here, read about peoples experiences in the forum, and even run it on your Smoothie. Helpfully this'll get me by. XLink KIA is avalible for Linux, Win32 and OSX. Maybe a Solaris version needs to happen. :)

Nothing really sucks more than having to get through a work day with a hot new game in your car just waiting to be played. Gran Turismo 4 finally released yesterday and because I didn't have a pre-order I wasn't able to pick it up untill today. So, expect that I might be sorta distracted for awhile. I can't wait to get back in the "seat" of my finely tuned Opel Astra (Team Pheonix) and start winning races. I was a big fan of GT3 and especially got into a competition with my brother-in-law, trying to get the best tuned car and duke it out on Leguna Seca. Now with network play (sorta) we will hopefully be able to compete without having to wait for holidays were he comes up here or we go down there (he lives in LA). After so many years of waiting its about time I got back into it. Besides, I haven't played games much at all in the last year... I used to be a big Final Fantasy fan (IX ruled, X blew, X-2 was a joke) with an undieing love for Final Fantasy Tactics (sadly the laws system in FFTA just killed the fun of the game), I played Disgaea for awhile but didn't have time for it. RPG's just are way too time consuming for me these days, so GT and the occasional DOA match are about all I have time for. Hopefully when FFXII releases it'll suck so bad that I won't feel compelled to play it.

I love old films. If you've never watch the old classics your seriously missing out. Tam and I have blown whole weekends thanks to our favorite channel Turner Classic Movies, which took the title acouple years ago when AMC went down the tubes... basically, good movies follow Robert Osborn. About 1970 everything just sort of fell apart in Hollywood and we're stuck in the rut to this day. Anyway, I've always considered my top 3 favorite films to be: 1) Blade Runner, 2) THX-1138, and 3) City of the Lost Children (the 1990's french remake)... but really, if you want some damned good films, you gotta look way way back, so if your looking for a good flick some day, here is a list of some films you really should check out that you might not have heard of:

• "You can't take it with you": I've been lucky enough to catch it 3 or 4 times and damn its a good flick. Lionel Barrymore is an amazing actor.
• "The Philadelphia Story": The cast of a century with the best first minute of any film, ever.
• "Harvey": Probly the best Jimey Stewart film featuring him as the lead.
• "Notorious": The best Hitchcock film. Period.
• "The Bells of St. Mary's": Ingrid and my main man Bing.
• "Adam's Rib": One of my favorite films of all time, the relationship between Kathrine and Spencer is amazing.
• "Desk Set": Required geek flick! The cultural shock of IBM's computers entering the working world in 1957, set in one of the best romantic comedies of all time.
• "The Thin Man": The whole series of films were awesome. William Powell is the king of kool, and Myrna Loy is awesome.
• "My Man Godfrey": A damn funny ass film.
• "Love Crazy": The best William Powell and Myrna Loy team up and one of the funniest flicks around.
• "City Lights": Chaplin's best feature film, possibly both the funniest and most romantic film of all time, silent or talkie.
• "An Affair to Remember": First half of the film is one of the best comedy team ups ever... second half just rips out your heart. (I turn it off half way through)
• "Arsenic and Old Lace": You just gotta see it.... most twisted comedy ever.
• "The Scarlet Pimpernel": Leslie Howard rules! The best film you've never heard of.

Damnit, there are dozens more. Chaplin, Clark Gable, Leslie Howard, William Powell, Myrna Loy, Cary Grant, Ingrid Bergman, Kathrin Hepburn, Spencer Tracy, Jimmy Stewart, Marilyn Monroe, Fred Astaire... all amazing. As you can see I tend to prefer romantic comedies because they are the light hearted with the right blend of wit and humor without all the slapstick that just gets boring. The dialog is what I love. lines snapping back and forth like blades in perfect balance, all seemlessly. Watch any of the films above (minus "City Lights" of course) to see what I mean about kickass and witty dialog. Anyway, modern films suck..... and color is just a stupid fad, imho. If you want to rent a good flick and cuddle on the couch with your lady while watching something really really funny, take this list and enjoy.

This isn't a pretty time for Sun, last week we saw more than 3,000 employees downsized. Most of these were in the UK, with word that at least one complete facility is being closed. I deeply feel for all those on staff right now. I talked with several people at Sun last week who were afraid that they were next and moral around Sun is taking a hefty beating, which is a shame after everyone started getting excited about the progress they've been making. This is a massively complex and frustrating problem for Sun. Just step back and think about this, on one hand consider for just a second Sun, what it does, what it provides, where they're going.... ok, now, on the other hand put 32,000 employees. They just don't balance out. Just here in the Silicon Valley I drive around at night when Nova can't sleep and visit several Sun sites and just looking at the number of buildings alone its hard to believe that Sun needs all these people. Sun hired way way way too many people and we're paying for it now.

Being a zealot I tend to wear Sun logo shirts about 90% of the time and at least once a week I meet someone that used to work for Sun... when I go to the kart track, when I go to the hardware store (almost all the local hardware stores are employeeing ex-Sun people it seems), when I got to the supermarket. You can't look anywhere in the Bay Area without finding people who used to work for Sun... and whats even more sad is that its not going to stop. From a business stand point Sun should realistically have an employee base of between 15,000 and 20,000. Prior to this layoff we had a count of approx. 32,000, and with this 3,000 person layoff we're down to 29,000. We've got a long way to go. And this is the hard part, when you look at the annual reports, company financials, or even the website and new releases its somehow inevitable to realize that layoffs are coming and they are going to be really really deep... and lets not forget that Sun's already gone through several really deep cuts in the past (including a massive and painful cut that was supposed to be the downsizing equivelent to ripping the band-aid off quickly). But, and this is a mother-f**king big ass but, when you look up from the financials and see an empty chair you can't avoid thinking about the people, mom's, dad's, geeks, engineers, managers, and all the individuals themselves who are going to get the axe. Who one earth has the heart to look at a Sun employee and say "I'm sorry, you've got to go for the good of the company"? I certainly don't. The words "for the good of the company" are only so comforting, and when the bitterness hits you after hearing "I'm sorry" the rest just sorta slaps you around. I don't want anyone to be released... but the numbers just can't add up untill you balance the equation.

So if Sun does in fact cut back to 20,000 employees are we going to be in the clear? My personal experience to date says no. Here is the problem, the powers that be don't seem to actually understand what anyone actually does on an individual basis, which means that unbelievable, essential, and visionary people are being annexed in these big cuts that are the lifeblood of the company. If you have a group of say 100 people and the company decides that they need to terminate the group, do you simply fire everyone? Hell no, you evaluate carefully which individuals are stars that need to be retained and which individuals are just sort of floating along... in a group of 100 you more than likely have at least 5 folks in that group that are really good and maybe 2 of them are stars, so you keep that 5 and then release the rest. You then take that bunch of 5 and put them in other groups where their talents can be exploited more fully. But Sun hasn't seemed to operate this way, they just amputate the whole damned arm and worry about it later. Anyone who's consulted for Sun in the past knows about this first hand, when every 6 months or so they terminate any consultant that doesn't have a vital justification, which is probly a good idea for cost savings... except that a week later they have to re-populate many if not most of those positions with consultants again, repeat cycle. And what does this point to? A faulty management structure made up of managers that don't understand the people underneith them. Is Jonathan to blame for John Doe getting the boot when he was a major asset? No, but John Doe's manager should have had the sense to realize that John was a much needed asset and retained him. And if these managers are doing this, then we need to look up the management structure another rung and figure out whats going on there. Internal communication and management structure are the keys here, and signs are that these are long time major week points for the company. Possibly because of poor organization. Possibly because of excess girth. Either way you slice it my gut tells me that so many essential persons at Sun will get chopped unintentionally if deep cuts brought us down to the 15,000 mark that Sun would hardly be able to function, which would mean hiring in a slew of people to compensate for the loss of a small group of visionaries.

This isn't the first time I've said this..... and I'm quite sure that it doesn't come as a shock to anyone. The patterns thus far are such that I wouldn't be supprised if Gosling himself was handed a pink slip, the lack of discression is that bad. And this isn't just about this current group of 3,000, this has been going on for a long long time. I'm sure we all know a handful of people who could leave Sun and not be missed, and many of those individuals probly wouldn't be shocked either, but we all also know at least a handful of really amazing people at Sun and we'd feel comfortable knowing that they were secure. Where, when, and how the remainder of the needed cuts will come is anyones guess. I'll sadly admit that despite all of Sun's amazing progress toward a brighter future that if they terminate another 10,000 employees without careful person-by-person analysis and then at the same time begin employing more people in India and China that Sun will be lost. Sun's coming to a cross-road, and its not a technical one... this is going to serve as a test for the future. No company of this global importance should have a future so uncertain as this when the possiblities are so wide. I urge the management of my beloved Sun Microsystems: proceed with caution and clarity.

Grids... Virtual Storage... Virtual Servers... Abstraction and commiditization while also building an enviroment that better consolidates and utilizes existing resources is a big deal. We've seen virtualization move alot in the last couple years. Server's can be cut into peices in two primary ways: Hardware Partitioning and Virtual Servers. In either case your carving up a server into smaller bit size peices and running independant operating enviroments on top. Just as a refresher, what are the options today:

• IBM: LPARS
• HP: nPars for partitioning and vPars for virtual servers
• Linux: UML, VServers, and Xen
• Sun: Dynamic System Domains (DSD) for partitioning, Solaris Zones for virtual servers

So why isn't VMWare on that list? Well, it is alittle more flexable not being tied to a vendor or OS per se, but I'll admit that its coming of age thanks to VMWare ESX. In the days gone past people would suggest that VMWare was an alternative to vPars or LPARS, and frankly it just wasn't... VMWare prior to ESX was a nifty way to run Windows on Linux or Solaris, or multiple contained operating systems for testing, QA, development, etc. Anyone who would have used VMWare to run production servers was just smokin' the wrong type of leaf, but all that changed with ESX, where we no longer simply run the guest operating system as a windowed app (VMWare Workstation) or as a system service (VMWare GSX). Now we introduce the Hypervisor, which is essentually the primary installed operating system that manages the various guest operating systems, in a sense replacing the functionality that a system controller would provide on a Sun or HP system. VMWare ESX now comes toe-to-toe with Xen. And we've recently seen that several top teir players are getting on the Xen train including IBM who's working on security, Red Hat and Novell that are looking at including it as future functionality in their products and even Sun is looking at it.

But while everyone looks strongly at the future of Xen... I'd like to remind people to not take their eyes off the Linux VServer Project. VServer is pretty slick, I definately like it. Think of it was providing what everyone wanted User-Mode Linux to be. VServer is pretty robust and really kool. I don't think its nearly as solid an implementation as Solaris Zones nor does it have such low overhead, but its damned close. Read up on VServer's benifits and consider it for yourself.

I started writting a paper on server virtualization awhile back but dropped it part way through due to my getting so bogged down in research and play that I had not time to write about it. If anyone is really interested in a solid overview of where all this stuff stands and where its going let me know and I'll consider updating and reviving the paper.

This evening Tam, Nova, and I set out to hit the Cambell British Food Centre to stock up on Tetley (British Black tea, we can't stand american Lipton) and Milkybar's (best chocolate bar ever, we American's just can't make candy like the Brits) and I took the chance to stop by Cycle Gear of San Jose on the way back so that I could scope out helmets for karting. I'm slowly building up my own collection of karting safety gear so that when I start racing on outdoor tracks where they don't supply you with the stuff I'll be ready. Mind you, I only went to look... but, damnit, they were having a 1 day 25% off sale. So, sadly, I found just exactly what I wanted and wasn't strong enough to say no. It didn't help me any that Tamarah was encouraging me to get it too...

My newest toy, an Arai RX-7 Corsair helmet with the Haga replica paintjob. Its a damn nice helmet indeed and hopefully will last me for many many many years to come (depending on how long the World Karting Association allows Snell M-2000 helmets). Most wives won't let their husbands do anything... mine presures me into splurging on my toys (within reason). Pretty kool indeed. Now if SHIFT would just release the Jet TI suits!

Apparently this was news back in November, but I just found the pictures: Ashton Kutcher has joined the l337 kilted krew to join Patrick of the ScreenSavers, myself, and a host of others. All the comfort of a proper traditional Scottish kilt, but without the weight, heat, and featuring pockets. Check him out sporting a black Workman's UltiliKilt (I love mine):

Best of all they started making my favorite kilt again! Time to stock up. If you still haven't figured out why kilts are superior to pants read the Top 10 Reasons to wear a UtiliKilt.

Damnit! I just can't take this any more...... holy shit. Steven Vaughan-Nichols (aka: SVN) writes yet another crackpot pile of shit entitled OSI Should Close Open-Source Licenses. Has this guy written a damned line of open source code in his life?!? He clearly just doesn't frickin' get it. "OK? OK!". Damnit. Choice you bastard... thats what its about. Not every frickin' project falls under the guidence of BSD, GPL, or MPL. The whole damned point of the OSI's license approvals are to measure the licenses against the Open Source Definition, in this way they are serving a function similar to the United States Supreme Court... the purpose of the organization is not to legislate from the bench, only to interpret and approve or disapprove. Perhaps it is time for the OSI to step up its role a bit and start laying down some law, but they'll need to do it carefully. The fact that there are 50+ licenses is a sign of two things: 1) No license can handle everyones needs, and 2) The open source world is a large and diverse place. Now, apparently SVN is using Redmond logic and would move to toss us all into a hole dictating what licenses we should us. Would it be nice to have <6 licenses that we all business and developer alike agree to? Sure. But it ain't gonna happen..... no way no day. The GPL 3 is gonna try its very best but there is a long hard fight ahead to pass it. His references to good ol' FInk help put his misguided faith into perspective. Open Source and Free software isn't something you can frickin' regulate. No one is telling me which licenses I can or can't use. So, fine, you rip away all but say 10 licenses from the OSI.... will it make a diffrence? Hell no.... there will be all these licenses floating around that won't have been managed or moderated by a singular body, and won't that be much better? No. This guy has lost any connection he may have ever had with the community itself..... maybe he never got it to begin with, just another asshole at LinuxWorld waving the open source flag without having any real idea of what this is all about. I say we hang this guy from his toes and stick him in a room with Richard Stallman for 72 hours with hopes that he comes out with an education. I'd suggest that the editors of eWeek find themselves someone with a clue in the very very near future.

Last week I read Georg's blog entry about the book by Tim Sanders, Yahoo! Chief Solutions Officer, entitled Love is the Killer App. Based on Georg's comments and other reviews I read after googling the book I got the feeling that perhaps ol' Tim here was simply taking open source and UNIX development philosophies and cultural methodologies and then applying them to the business world. I was interested in this, so I picked up the book prior to my leaving for Boston and read it on the plane. Now, I'm a slowwwwwww ass reader... I rarely read non-technical books because I read so slowly that I don't even bother to read many of the classics that I'd like to; for instance, I've never actually read "The Hitchhickers Guide to the Galaxy" or "Lord of the Rings", it would just take me months to get through them. I can blaze through technical books though. Anyway, Tim's book is really light, double spaced and a quick read, so I actually finished it during my 7 hours of flight time.

Okey, so the book wasn't really what I expected. His philosophy is pretty simple and somewhat common to the UNIX philosophy but only in an abstract and selfish way. His system breaks down into 3 parts:

1. Learn everything you can, read alot, only buy hardbound books because they make you look smarter, and mark the living shit outta 'em.
2. Network all you can. Share your knowlage, connect your connections with other connections without agenda, become a facilitator of goodwill by bringing people together.
3. Be compassionate, cut others slack with the hope that they'll cut you some, honestly be interested in what people think about your advice and learn how well your networking efforts and knowladge are working. Essentially, get feedback on and fine tune steps 1 and 2.

So the paralells of the UNIX/open source philosophies are there, to some extent. Be kind, help others, learn everything you can, try to pass that info along as best you can, and accept all the input you can so that you can get even better. But, while this is a stark contrast the business worlds long held "greed is good" philosophy in which you try to keep any advantage you have a secret so as to maintain an edge over your competition (whether for your business or for your job), this method only goes half way to acheiving the true UNIX/open source philosophy. To me, the essence of UNIX is in community, good will, sharing, caring, and contributing to make the world in some way a better place, whether for an individual (helping someone debug an app, or giving helpful custructive guidence) or for the whole of the community (writting a better [insert nifty app here], or better yet, documentation that helps you learn how to write that nifty app yourself). Tim's methods stop short of the UNIX way in that you don't actually contribute to the world, just your surroundings. People will get to think well of you as an individual because you help facilitate progress and interaction (what we formerly refered to as "a team player") yet your not really giving back to the world as a whole unless you put all that nfity information down on the page or otherwise make it avalible.

The, as he calls it, "lovecat way" (*cringe*) seems less to me like a business adaptation of the UNIX/free software philosophy than it does a modernized and personalized implementation of the Macy's business model of yor. "I'm sorry ma'am, we don't have that item, but Burdstroms should... ah, yes, and its on sale! May I call and set one aside for you?" In this way you familiarize yourself with your field, keep as current as possible, connect (network) parties that would benifit from each other without expecting anything in return, and constantly ensure that you are helping the relationships your involved with whether by being active or by knowing when to leave them alone so as to not pester. The idea here is that by giving without expecting return you build a good reputation and generally feel much happier as a person, building others up rather than ripping them down. But, this is a business book, right? So of course, what your doing is figuring that if you do 6 good deeds and only 1of those benifit you then you came out a winner in the end anyway and over time your average should improve. This is akin to always buying people nice gifts even though you get shitty ones, in hopes that one day they'll actually get you something nice. I dunno..... its all interesting and whatnot but not what I was hoping for. I feel that you should strive to benifit the community as a whole in a direct way by contributing to efforts, or by helping others to contribute themselves in a varity of ways, such as being a cheerleader, an organizer, a writter, a designer, a coder, an evangelist, a supporter, so on and so forth. His scope is too small..... but then again, at least its maybe a start for some people.

If you like business books then buy a copy and flip through it but otherwise don't waste your time.

I'm down at our (my employeers, not cuddletech's) colocation facility today doing NetApp installs in preperation for a nearine replication setup I'm putting together. While I wait for floppies to load I thought I'd walk around the server room and see what folks are running these days. Obviously I can't say which colo, nor which companies are hosted here... but it's the silicon valley, so use your imagination. I guarrentee sites that you often use are located in this room.

So, the far and away leader in this server room is Sun, with the bulk of installed systems. There are over 40 diffrent cages in this room that are populated, so its a good mix of diffrent businesses. Running a close second is Dell, the number one X86 vendor here. Third would be HP/Compaq X86 systems. Behind that you have HP PA-RISC systems and other manufacturers.

When it comes to networking, the leader is of course Cisco. There are a growing number of SonicWall Firewalls installed, but still the dominate solution is Cisco PIX.

We have one cage here that is nothing but a row of 6 EMC Symmetrix, but they are the only ones. Several cages are using HDS Thunder and Lightning. Storage from Sun, Dell, and HP/Compaq follow the same trend as the servers, obviously sold in bundles.

The only real shocker here is that when I walked in today there was a new occupent in a previously empty cage, and whats installed? To my supprise 2 racks of IBM pSeries systems. I'm not sure if they're running Linux or AIX, but it took me by suprise. Believe it or not, IBM isn't a commonly found vendor in most Silicon Valley colocation facilities (except maybe IBM's colo south of San Jose).

Anyway, maybe thats interesting, maybe not... if nothing, this report has helped pass alittle time. :)

I just learned of an interesting scripting language today: Lua ["LOO-ah"]. Apparently its becoming popular because of its use in several games including "World of Warcraft" in which apparently users can customize their charrectors and UI. It boasts nice C interoperability and embedding capability. For a quick overview check out the wikipedia entry or read "Programming with Lua" online. For thoughs who want a fast track to digging into it, just skip right over to the Lua reference manual. Gentoo and Blastwave already have it, so merge or pkg-get it and give it a whirl.

I'm back from Boston. I went out there for 2 wiz-bang days to talk with press and analysts about OpenSolaris. I flew out Tuesday morning at 6am so that I could arrive that afternoon in preperation for a dinner with several folks covering our progress. I ran around the show in the hour I had to kill between arriving and the dinner, and all I can say is: samo samo. LinuxWorld is so damned commercial. Everyone was giving the normal corpie rundown. Novell's booth seems to get bigger each event, but this is the first non-California LinuxWorld that I've been at so perhaps they normally have a bigger presence. I strolled through the .Org Pavillion area where the REAL open source projects where and had a nice chat with the PostgreSQL guys and generally mulled around.

So the dinner itself was interesting. On the external side was myself, Steven Christensen of SunFreeware Fame, and Jed Dobson from Dartmouth. Steven is a super kool guy and knows more about Mathmatica than you ever thought someone could. Meeting him was a real honor. On the Sun side was Andy Tucker (Distinguished Engineer working on various parts of the Solaris Kernel), Claire Giordano (the force behind the CDDL), and Tom Goguen (Vice President, Software Group). Jessica Bookach and the rest of the PR team were there coordinating. On the press side we had Scot Peterson of eWeek, Steven Shankland of CNet, Hiawatha Bray of the Boston Globe, and several others who's names I won't disclose. Overall it went well, diverging often onto a variety of tech subjects but always trying to come back to OpenSolaris. This was the first time I'd talked directly with analysts (knowingly anyway) and I was supprised and how easy going and funny they were. Steven was really kool, extremely knowladgable about the industry as a whole, and fun to talk with. Scot is an interesting guy, very direct, takes no crap, and has a way of ripping away all the fluff to leave with the point... I like Scot, he's definatley my kind of guy, don't pass up a chance to meet him. Now Hiawatha, wow! This guys picture should be in the dictionary next to "personality", he's a warm, personalable guy, with charm from here to next week. If your having a tech event in the Boston area and don't want it to be boring, invite Hiawatha.

The bulk of the dinner I begged the staff for more coffee and chatted with Steven. Eventually Andy came over and joined in. I'm not entirely sure how much Steven got out of it because we were moving from topic to topic, discussing a wide variety of things reguarding open source, licenses, community, OpenSolaris, Linux, Red Hat, IBM, HP, and on and on. I probly should have been more focused, but I'm interested in the big picture of the industry and Sun's role to play so I can tend to chat on just about any aspect of it with the notable exception of the Java side of the house. Anyway, I did the best that I could. Like I said, I could have been more focused, so I'll have to work on that.

After the dinner everyone split up, I talked alittle with Tom about how he's adjusting to life at Sun and we talked alittle about Carly getting canned, but nothing major. He took off and then I, having nothing to do and being alone, went and explored. I used to live in Concord, New Hampshire, so I had alittle familiarity with the place, but the trick is that when I lived there I was probly between 5-7 years old before my parents returned to California. Because of my brief stay in New England during my youth during my tender early years I identify alot with the place, namely my affinity for cold weather, good clam chowder, lobster, and the architecture of New England. I went to my first baseball game (that I remember) at Fenway, so I'm a Red Sox fan, that sorta thing. Anyway, I hit an Irish pub and had a pint of Guinness and then walked around for about 2 hours, stopping for coffee and donuts at one of the "Dunkin' Donuts" that you find on damned near every corner of the town. After I got tired of aimlessly walking I went back to the hotel, watched CSPAN and the local news for awhile, called Tam and talked for about 40 minutes and then crashed out.

The next morning I had lunch with the Sun folks and the other pilot members present. It was a nice lunch, we talked about this and that, but sadly I was yet again unfocused and left the lunch wishing I'd asked this or that. There just wasn't enough time. After that was over, again, everyone split up and I hit the LinuxWorld floor for about an hour before getting bored stiff. I talked with IBM about some details I couldn't get reguarding LPARs control from an HMC, and talked with all the Infiniband vendors there, chatted with Oracle alittle, got into an argument with LSI Logic, made fun of the HP guys, and then trolled the Sun booth for awhile. The funny thing about the Sun booth is that, 1) I had more Sun logos on my person than anyone from Sun (in fact during dinner I was the ONLY person wearing a Sun shirt, and I was wearing 2!), and 2) No one knows who I am. I broke out the kilt for the second day and most of the folks at the Sun booth just thought I was a wierdo. So I left and wandered around Boston alittle to take some pictures. I can tell you this, no one in Boston has ever seen a kilt before, little less a camo kilt. I got lots of laughs, snickers, and comments behind my back, but then I'm used to that. After about an hour of walking around I was dieing of bordom, lonely, and missed my wife and daughter. I just don't do well away from either of them at all, so I found the nearest taxi and went to the airport. I got to the airport 3 hours ahead of my flight, but there was nothing else to do in Boston, so I called Tam for a pick-me-up, talked for about 30 minutes and then trolled the airport untill it was time for the flight to leave. After being crammed into a coach seat for 7.5 hours I finally was back in the Bay Area and glad for it. I picked up a burger on the way home, hugged Tam for about 5 minutes and then we crashed out. And now I'm back at work. Remember, I don't work for Sun, right, so I had to take vacation to go out there.

So, all in all, the trip sucked, dinner was great, the press and analysts were awesome, meeting Claire, Andy, Tom, and others was great, the show sucked, Boston is beautiful but boring on foot, and flying coach is just lame. Hopefully next time something like this happens it'll be here in the Silicon Valley, since I love the chance to talk about Sun, Solaris, and particularly OpenSolaris but having to spend almost 10 hours traveling in cabs, shuttles, and planes each way just takes the fun out of a 3 hour event.

Wanna be a lab rat? If you take and pass your Solaris10 BETA certs they count for real... at no cost. Get 'em while you can. I'm already signed up. :)

How many geeks can honestly say that they have a family dedicated to open source just as much as you are yourself? Thankfully, I can. I was just downloading pictures from my digital cam that I took over the weekend and didn't realize it but both Nova and Tamarah just so happened to be wearing their Enlightenment shirts. Dedication. If only more coders were as passionate as they.

Some nights just aren't your night. The Sunday Night Heats just didn't go my way at all. During the first heat I had a bad kart and slipped from P1 to finish P2. Then in the second heat I had a great battle but we all a ran so tight that I got edged out when two drivers got overly agressive, forcing me to either do "the right thing" (show descretion and conceed the corner) or be an asshole (hold the line and run the jerk into a wall)... I did the right thing and paid for it finishing last (proverb anyone?). In the third heat the shit hit the fan, a 110lbs 12 year old kid was blocking me in P4. Now, the problem with little kids in karts is that because they are so light they might not be able to corner well but they can fly in the straights which means your all over the back of 'em throughout the chicanes but on the straight they just pull away. So I'm all over the back of this kid but I didn't have quiet enough speed to overtake, so I said for a mistake... then, he screws up, hits a tirewall hard and gets under the tires, but I'm 4 inches off his rear and my kart has almost no breaks, so I hit him, which actually helped because I knocked him free from the tirewall and off he went. Trouble is that the marshalls knew he was holding me up and thought I slammed him, even though they know its not my style. Black flag comes out, meaning I'm in trouble and have to pit for "a talk". I explain and they send me out, but at that point I'm 2 laps down and can't make it up. The truoble is that I could have made it up to P1 if I'd be allowed to stay on track, but the black flag killed me. They realized the made a mistake, and gave me back P4 on the points sheet, but I needed a P1 bad due to my earlier P6 shutout. In the semifinal I had a kart with a bad engine. The trouble was that the engine was really fast but it took forever to get it up to speed, so if I could run flat out without having to break I was fine, but when I came into traffic I'd get killed when I'd break because I just cuoldn't accellorate to stay with 'em. This all meant that I was either: a) not going to advance from P6, or b) going to need to hit someone hard to make a pass. I thought about it for 2-3 laps and finally just pulled into the pits mid race and gave up on it. At that point I would feel better with a DNF (Did not finish) than two last place finishes. So, for the first time in several weeks I wasn't able to make it into the final event (top 10 point scorers). A totally aggrivating and upsetting evening... but, as the saying goes, "thats racing"(tm). We'll see if I can finally get a good string of luck next week.

Continueing on our journey of handy dandy Solaris features I thought tonight we'd examine a handy feature of Solaris that even many long time Solaris users aren't aware of: IP Multipathing, or simply IPMP. I found out recently that people in the Win32 world call this "teaming" but in the UNIX world we typically just say "adapter failover". With multiple interfaces on a Solaris box we can provide some extra bandwidth, which we call trunking, or we can use them as redundant links by way of IPMP. While IPMP is all about adapter failover, generally its not really the adapter your worried about, its the switch that its connected to. If your enviroment has redundant switches and you are either sick of your network admin having to upgrade firmware every other week or you just don't trust 'em, then IPMP is for you. You can read all about IPMP in the "IP Services" manual but I'll skip past all that cruft and give it to you more easily.

IPMP's job is to detect and react to network failures, this can be accomplished in 2 ways:

1. Link Based Failure Detection: In this case the IPMP daemon (mpathd) watchs for standard "Link Up"/"Link Down" reports from the adapter. If the an adapter reports "Link Down" it moves the IP configuration for that interface to another adapter in the same group that has a link.
2. Probe Based Failure Detection: In this case the IPMP daemon (mpathd) actually sends ICMP packets back and forth across the network (ping test) to ensure that connectivity actually exists. If a packets stop coming the adapter is failed over.

Link-based detection is simple to setup and adequate for most needs. Probe-based failure has obvious benefits but its more complex to setup and administer and therefore we'll just discuss Link-Based. The setup is really simple, you simply add two or more interfaces to an IPMP group and your done. Because IPMP uses ifconfig for all its interaction you can do this simply:

root@betty export$ ifconfig -a
lo0: flags=2001000849 mtu 8232 index 1
        inet 127.0.0.1 netmask ff000000
hme0: flags=1000843 mtu 1500 index 4
        inet 10.0.0.70 netmask ffffff00 broadcast 10.0.0.255
        ether 8:0:20:aa:d3:e4
root@betty export$ ifconfig hme0 group grp1
root@betty export$ ifconfig hme1 plumb
root@betty export$ ifconfig hme1 group grp1
root@betty export$ ifconfig -a
lo0: flags=2001000849 mtu 8232 index 1
        inet 127.0.0.1 netmask ff000000
hme0: flags=1000843 mtu 1500 index 4
        inet 10.0.0.70 netmask ffffff00 broadcast 10.0.0.255
        groupname grp1
        ether 8:0:20:aa:d3:e4
hme1: flags=1000842
 mtu 1500 index 5

See? Easy. But like all things that you do with ifconfig, this all gets blown away at reboot. To make these changes persistant across reboots we can add the group arguments to the /etc/hostname.(interface) files. The contents of these files are simply arguments passed to ifconfig when the interfaces are configured, so treat it appropriately. In the following example I show both of my hostname files, in this case hme0 and hme1:

benr@betty ~$ cat /etc/hostname.hme0 
betty netmask + broadcast + group grp1 up
benr@betty ~$ cat /etc/hostname.hme1 
group grp1 up

When you see a "+" in a hostname file its simply going to be replaced by the appropriate information. The full expansion of the hostname.hme0 line when run will actually be "ifconfig hme0 10.0.0.70 netmask 255.255.255.0 broadcast 10.0.0.255 group grp1 up". You can see that the host name "betty" is replaced with its IP found in /etc/hosts, and the +'s are replaced with information from /etc/netmasks.

One last thing you want to do is "touch /etc/notrouter". This will keep Solaris from trying to route packets, and is primarily a safety precaution.

Once IPMP is setup go and test it! Here is what my Linux box sees when I yank the active adapters ethernet out of it during a ping test:

64 bytes from betty (10.0.0.70): icmp_seq=8 ttl=255 time=0.523 ms
64 bytes from betty (10.0.0.70): icmp_seq=9 ttl=255 time=1.44 ms 
64 bytes from betty (10.0.0.70): icmp_seq=10 ttl=255 time=0.521 ms
64 bytes from betty (10.0.0.70): icmp_seq=11 ttl=255 time=0.526 ms

Not bad, that 1.44ms lag is the failover. You can failback and forth as quick as you like. By default when you "fix" the origonal adapter IPMP will fail back onto that interface keeping things nice and pretty. You can tell IPMP to not failback to the origonal by changing the parameters in /etc/default/mpathd configuration file.

For details on using Probe-Based failure detection or more details refer to the IPMP section of the "System Administration Guide: IP Services" manual. IPMP has been a feature of Solaris since Solaris8 (10/00), however significant new functionality exists in Solaris10.

The other day I noticed Alan mentioned the release of 6.8.2 I decided to try it again. I'd used Xorg when the Composite first was integrated but it was less than perfect, but never the less kool. We could debate all day the virtues and shortcomings of Xrender and Composites implementation but so far its the best we've seen. Thanks to Gentoo I simply yanked XFree86 (unmerge) and merged in Xorg 6.8.2. Once you do that, to really get the Composite experience I upgraded to Enlightenment DR16.8 from CVS (not yet released, still in testing... shhhhh), damnnnnnnnn its nice. Kim Woelders, who took over DR16 development almost 2 years ago has just been doing unbelievable work. With DR16.8 you can enable composite from within the window manager and alter transparency on a per window basis using a window property (no more xtransset!). Its ohhhh so smooth.

The fullsize shot is alittle grainy due to JPEG compression, sorry about that, but as you can see in the shot Composite may have its downpoints, but its damned sexy and DR16.8 lets you fully harness it.

Once the excitement of the OpenSolaris preview settled down I was able to return to other projects, chief amoung them has been a book on the Oracle Database written for sysadmins, entitled "The Cuddletech SAs Guide to Oracle". This book was written at the begining of November in just a couple days. Its a journey covering the aspects of Oracle that an SA actually cares about with out all the baby talk or pointless bits. The thing started as notes, then became an article, and then a book. It's about 100 pages long right now. In mid December SAGE put me under contract to have this made into a book for them and I'm trying to speed through the review and editing stages, but I could really use some help from you guys! I would love to get feedback on it from other SAs both experienced with Oracle and those who've wanted to learn it for years but haven't learned it as well as they'd like. Just about anyone is going to like the RMAN chapter. And I sum up alot of Oracle odds and ends of interest to people in the last couple chapters. I think its pretty good, but if any of you can help me make it better I'd appreciate it.

The book was written in LaTeX and is avalible in two forms: Pretty PDF and not so pretty HTML. If you can, check out the PDF, its much kooler. If you have questions, comments, corrections, complaints, whatever, please mail them to me. I want to produce the best book possible to help other SAs who could benifit from it, and that means everyone pitching in where possible, as is the UNIX way. Thank You!

Over at the CBS Marketwatch site the top headline popped up (might be gone when you see this): DELL DOWN AFTER BELL, the abstract on the front page continues " The No. 1 PC maker posts higher revenue, but net earnings slip and sales miss most aggressive targets." Once you read the article you find that Dell actually did really well, but they got slapped with a $280 million one-time tax bill. So, in reality Dell was up solidly and grew the business... but does this warrent such a grim front page headline? Don't get me wrong, Dell can burn in hell for all I care, but I know I hate this sorta stuff when it hits Sun, SGI or Cray. I probly take these things too personally, but hey, someones gotta.

On an somewhat-related sidenote, I had my first ever meeting with Dell yesterday. Two of their sales guys came over to the office and gave me a (creepy sounding) "Customer Touch", a meeting in which we talked about Dell|EMC. I've been spec'ing out a new storage solution and as part of my due-dilegence wanted to at least get quotes and info from EMC about CLARiiON. Sadly however, EMC's sales force sucks f***ing ass! Damned incompetent fools can't even give me a damned quote and took a week just to return my call! So, I gave up on EMC's shitty sales people and called Dell, since they resell EMC product. Dell was much better about getting me the info I needed and even arrange this meeting to come talk to me. Anyway, I just had to ask some questions to these guys, like "Tired of the 'Dude, your going to jail!' jokes?", to which the told me about an interesting meeting in which Michael Dell addressed the marketing team and supposedly said "Dude, what a mistake!". I also asked about the "Dell Tech Force" advertising (seen during Michael's Keynote at OracleWorld) and the fact that having my systems vendor having 20 diffrent partnerships is really concerning when I can choose full featured vendors who can supply more or all of what I need, like IBM and Sun. They simply talked about the normal "best of breed, providing flexablity and oppertunity for our customers" crap. Sure, they have a point, but I want a systems company, not a VAR who slaps their badge on everything they sell, which is essentially what Dell comes off as sometimes. And, of course I just had to grill 'em on their constant misuse of the term "Industry Standard". Intel is not industry standard. HTTP is industry standard. Fibre Channel is industry standard. Win32 is not industry standard. Take any instance in which they way "industry standard" and just replace it with "most popular" and you'll get a more accurate picture. Oh well. So be it. I hate Dell with a passion, but then they have done a good job, to my eyes anyway, of being the big player that plays nice and largely by the rules. When Grandma, Uncle Ted, Small and Med Business can all rely on a single vendor for lost cost solutions, they have gotta be doing something moderately right, especially when they compete based on price rather than pushing customers around like a bully that you can't/won't stand up to (Microsoft). Oh well.

Mr Greg Papadopoulos has stepped into the big spotlight and started a blog. I'm glad to see this because I always hear various Sun folk mention him but don't know anything about him or ever hear from him during Sun events. I'm a small-fry, so all I get is what we all get in webcasts and stuff. I had no idea that he was in touch with the community or license issues, etc... very kool to see.

So if you slacked off like me and keep hearing Greg's name but aren't really sure who he is, here's the skinny. He's Sun's CTO, taking over after Bill Joy left the slot. He's been at Sun since 1994 serving in a variety of positions including VP of Technology, so he certainly knows the job. So, if you think he's got no cred, check this, he was Senior Architect and Director of Product Strategy for Thinking Machines "where he led the design of the CM5 massively parallel supercomputer". Okey, so maybe our beloved Thinking Machines died in a flaming ball of d00m, but hey, he's got an immediate Karma +5, Stamina +10, l337 +40 in my book. So, the guy has been around the block acouple times, seen good times, seen bad times, and slugged through them as best as he could in a variety of diffrent positions throughout the organizations. Very kool and useful indeed.

His update today (his second post) was interesting. I don't know if I completely agree on his considering tossing the term "the community" around as being Cathedral like (I don't like Cathedral and the Bazaar quotes anyway), simply because every member of the community should wrap themselves in it. What makes the "community" so interesting and helpful? You download the Linux kernel, you install it, you chat with some folks in IRC and then you are free to say "We're looking forward to the NFS patches next month", using "we" to mean the community. You appoint yourself, you don't ask permission, or refer to the community in abstraction. I see this sort of problem at Sun actually, internally, if you refer to your own organization in the 3rd Person you need to step back, I think, and say "Why don't I feel like I'm an integral part of the organization?" Communication is key, which is why ever employee, or community member, or even groupy, needs to feel a personal connection with the whole, which isn't something you grant, but something that is inherent to the movement itself. ...Assuming that makes any sense. :)

I'm probly just hitting him with symantics... what I LOVE is the fact that he finally said what I and others have been saying for a long time: "A footnote on CDDL and GPL code co-mingling. The issue of license compatibility with co-mingled GPL and CDDL code is an issue with GPL license, not with CDDL." Amen brother! It's about time it was said plainly at Sun. Jonathan has pointed to it, but not just outright said it. I'm glad to have finally heard from the lips (well fingers, I guess) of Greg himself and look forward to his future blogs. Greg and Jonathan are really going to help drive Sun in the right direction and I hope that more and more of Sun's employees embrace, but even more so understand the direction their taking us all, internal or external.

I'm not getting anything useful done tonight. I'm having one of those uber-tangent nights where you look into one thing, which interestest you in another, and another, untill you've spent hours researching and exploring things of little to no real value beyond interest. So to try and salvage something out of my tanget night, I'll blog some of it. ;)

I started playing around with Google Maps today, which has been helpful in planning an upcoming trip to Boston (details later). The detail of the maps is in many ways superior to MapQuest, with the only exception being the denotion of one-way streets. Otherwise the maps are much clearer to interpret and definately to navigate by way of scrolling the map rather than clicking a border to move the map or clicking to recenter like MapQuest. Google's starting to become a big-scary company with too much influence, but damnit they do things so frickin' well that you just don't really care so much!

My last 2 hours of tangent have been spent on IO buses, namely InfiniBand, HIPPI, ESCON/FICON, and PCI Express. InfiniBand is something that I tend to poke my head into every now and again to see what progress is coming. Interestingly PCI Express becomes hard to really understand because its so diffrent than traditional PCI (or PCI-X), frankly I think we might have been better off renaming PCI Express to something else to differentiate it but there are enough compatability concerns that I guess "PCI Express" works as well as anything else. InfiniBand is going to really be interesting though, especially if it moves into the FSB, replacing PCI, HyperTransport, or FirePlane. The day when CPU, Memory, and IO are all interconnected internally and externally by Infiniband will be an interesting day indeed... if it ever happens. When you dig a bit you find that not only was Infiniband intended to be a high speed low latency flexable architecture interconnect (as currently used in clusters), but it can be utilized internal and external to the system ultimately with the intention to consolidate mutliple existing links and HBAs (Ethernet, Fibre Channel, etc) with a single interface... imagine a high performance system with only 2 cables into the thing: power and InfiniBand! And not only that, but making every device in every machine accessable to other systems on the fabric, complete with RDMA (Remote Direct Memory Access). In years past I've viewed InfiniBand as a scary technological disaster to consume and replace every bus on the planet, but more and more I'm warming up to it... moving at 10Gbit as it now can, there is no reason to fear bottlenecks that would be present with putting your ethernet and fibre channel traffic on a single link... now the problem is finding a CPU that can actually drive the thing, even with offloading on the HCA (Host Channel Adapters, the InfiniBand equivelent to an HBA) you still have so much throughput that just proccessing that ammoung of data could be a challenge... but then thats the point right? However, I'm afraid that if PCI Express and Infiniband stay on the current course we're just going to see PCI-E as next generation PCI and probly not even threaten HyperTransport or FirePlane for the FSB, and then watching InfiniBand simply becomes an alternative to Myrinet or even fibre channel. I dunno... it's all unclear. Especially as very interesting technologies like Sun FireLink (aka: Infineon PAROLI - Parallel Optical Link) have silently slipped into obsolecence as technologies like 4G fibre and 10G ethernet linger just around the corner. So many possibilities... so interesting... so exciting... and probly such a waste of time for me to research. :)

But, speaking of storage... I know you've seen this already but damnit striping Apple Shuffles is just frickin' kool! Looks like he's getting about 3MB/s on that setup, which by my calculations suggest that either: A) He isn't optomizing someplace in software, B) Taking a massive overhead hit by using all 4 Shuffles on a single USB2 link, or C) The storage device (CF?) in the Shuffles are increadably slow... which would probly fit with the pricetag frankly. Either way, a kool thing to do. I overheard someone (shhhhh) mention playing with ZFS on USB storage devices not long ago... stripped USB pens anyone? :)

Okey, now to see if I can actually get back to work on my book before my editor kills me..... frickin' 3 weeks trying to write 1 sentance!!! Shockingly sad, I know. I wrote a 100 page book on Oracle in 3 days but the first paragraph have been re-written 40+ times and still isn't working........ damnit, Rik's gonna kill me. If I could just get a good first paragraph the first chapter would be done in 2 hours and I'd finish editing in less than 8... damnit. I had the same problem with my SSP review, almost 3 weeks trying to write the first couple sentances with no luck, over and over and over, wasting hours and hours... and once I got that one paragraph done the review was done in 45 minutes. Damnit I hate writting. ;)

After their long slumber, the HP Board woke up and terminated Carly. This should have happened a long long time ago. I've even told Jonathan that the best thing Sun's had going for it is Carly Fiorina, who quarter after quarter finds new and imaginative ways to piss off and alienate its own customer base. HP-UX Ported to Itanium2..... uhhhhhh, ok. PA-RISC to be terminated..... uhhhhhh, that sucks, but it sucked anyway. DEC/AlphaWorks divisions trashed, lumped in with move to Itanium2 on HP-UX....... uhhhhhhhh, WTF? Carly has be so kind to us all by lifting up her own customers on a golden platter for Sun to take in.

So what does this mean now? Can HP get its act together? Can they salvage the ruins of the EV6 and Tru64 line? Can they possibly hope to shift customers back? All unknown. But what is clear is that Sun had better put its head down and hold on tight because we need to fight fast and furious with a resolve to protect these customers from being fscked over again.

I've finally rehashed and upgraded my forums. Sadly, however, it's a forklift upgrade and I'm moving things around a bit meaning that all of the posts in the old forums will be trapped there. You can find the new forum on the nav bar or directly at cuddletech.com/forums/. So why more forums? Sun already has 'em right? They do, but I thought it would be a good thing to have non-Sun run forums avalible to the community so that no one feels like their posts will suddenly disapear into the depths if they post something unpopular. Feel free to discuss, debate, and learn more from each other there.

Other forums existing in the same place for Solaris10, Storage (this consolidates my previous Fibre Channel and VxVM forums), Enlightenment, and Solaris Porting. You can register for an account or post anonyously. I want to keep the forums no-reg, but if the spam kings keep it up I might have to go for registered users only.

I was checking /. today and saw an add for this: the Mailfoundry Email Filtering Appliance. You'll notice that the appliance is oh so clearly and proudly a Sun V100. See? Now thats the kinda of marketing you want. Works for Solinus and works for Sun, and just knowing that it's a V100 makes me want to give it a shot. Very kool indeed.

Good ol' Ashlee Vance in Chicago of The Register is reporting on the suggestion that Sun should aquire Unify proposed by Paul Murphy of Linux Insider in his article Sun Should 'Unify' Open-Source Software. I've gotta say, I really have come to like Paul, of all the press he's probly the most centered of the journalists out there. Ashlee of course, has also done a great job of pulling everything out there together.

As we discussed earlier, I just can't see this being a good move. You can read up on Unify DataServer 8.2 on the Unify website. Paul makes a good argument for Unify, but I just don't think this is wise. Sun doesn't need yet another aquisition right now, and the open source database market doesn't need yet another player, imho. We need to put work and weight behind existing projects to build them up. Heck, despite my objections to Firebird I'd recommend it in a heartbeat as opposed to Unify. Sun has the engineering power and knowhow to really change the world for the better, and it can do that by really helping build PostgreSQL. Why pay for a donkey when you can jump on a stallion and ride into the Sunset (no pun). Besides, we could all use a little PR pickup and gettin' down and dirty with PostgreSQL is just the way to do it. Unify just adds anothing bothing peice of software to the middeware stack, and even if its open sourced under the CDDL (preferably) it's just going to be yet another database... why should anyone bother with it given its feature set and the fact that we've already got more than enough databases to contribute too? I just don't see Unify being a good move at all. If might not be so bad, I haven't used it yet, but we can get alot more gain at a much lower cost by working on the existing databases. Maybe I should order 100+ PostgreSQL tshirts from SourceWear and hand 'em out at the Menlo Park campus. :)

In my last blog entry reguarding the rumored SunDB I immediately discounted the Firebird Database as a contender. In the comments posted to that entry it was suggested that Firebird shouldn't be so quickly discarded and may even be a "diamond in the rough". Well, if I'm wrong I'm willing to explore it and admit that I'm wrong if thats the case, afterall, what if Firebird really that great? I certainly don't want to miss out on a good thing. So thanks to Gentoo I quickly emerged Firebird 1.5.2 and started playing.

Something that you learn about Firebird quickly is that the ties to Interbase are strong and there is seemingly a battle between maintaining Interbase compatability and taking the codebase in a new direction. I'm not sure what the result of this will be long term, but apparently there are way more people out there still maintaining Interbase databases than I realized. After only 5 minutes of digging through the website and docs you also realize that Firebird has a strong positioning for Win32, again most likely due to the Interbase/Borland lineage where it would have been commonly utilized with Delphi/Kylix on Win32. If your stuck on Win32 for some reason and are looking for a low-cost/no-cost database then Firebird may be just the ticket you've been looking for. Even with these ties, the *NIX ports are pretty solid. You definately can appreciate the slenderness of the RDBMS, with only 11 tools, all of which are pretty simple to understand and use, your not getting buried under a downpour of tools like you'd see with DB2, Oracle or even Sybase ASE. It's pretty easy to quickly dig your fingers in and go. The documentation in general needs a lot of work, and I'm almost tempted to help out, but this will come in time.

While I like several of the tools (fb_lock_print and gstat are kool) the lack of features avalible in more powerful databases becomes clear after some playing. Lack of scheme support and tablespaces can be avoided but for those spoiled by more full featured databases might miss it. Both of these features, however, are really unneccisary for small to medium sized databases, it's unlikely that you'll be running such a large database that you will be forced to ensure that the disk subsystems for indecies and tables are on diffrent disk structures, or that you'll need schemes to keep things seperated. SQL99 support is lacking, which PostgreSQL does have, but again for most small to medium size databases this isn't an issue either. Replication is a feature growing more and more important every day, but Firebird's replication offerings are primarily bring overs from the world of Interbase. Currently you can purchase IBReplicator from IBPheonix, but sadly you are required to use a Win32 workstation to manage the replication, although replication can occur between Win32, Linux, and Solaris systems. New replication projects like FiBRE: The FireBird Replication Engine are underway, but if you want replication specifically for Firebird your going to have to either pay or wait.

Everything I've seen thus far does indicate the Firebird is very very fast and efficient with resources, but I'll admit that I'm not building large enough databases to really tax the thing. So, at least for now, clearly if you want a small and light database with many of the features of mainstream RDBMS's and you don't like SQLite, then Firebird is definately to be considered. If you ever think you'll run your database with or on Win32, then hands down Firebird is for you. And, if you have wanted to get involved in database development and just haven't felt like your help is needed, then definately consider getting onboard with Firebird and help out.

So, Firebird definately has some high points. But when we consider what a major house like Sun might use for its own database, Firefird just doesn't fit the bill... at least, not right now. PostgreSQL is full featured enough to at least try to take on the big boys of Oracle, Sybase and DB2, while Firebird is getting there. If you look at Vulcan: The next major version of Firebird you'll see that they have some seriously kool stuff going on, and there is no doubt that other databases had better keep up or their going to find Firebird eating away mind share in the not to distant future, but not right now. I will admit that I feel like finally our lost brothers and sisters on Win32, sinking in the sea of C++ and M$ .Net terror finally have a really good option for a database, Firebird provides these guys with a really well built and Win32-happy array of development tools and interfaces that don't have that tacked-on feel that I'd imagine databases like MySQL and PostgreSQL have, and with out all the "Damnit, just run Linux" feel. I'm not fan of Win32 but we'll at least acknowlage that there are developers still stuck on that ancient plaform whether they like it or not.

So, at least in my book, I still think the 3 greatest databases are SQLite (small), PostgreSQL (med), and Oracle10g (large/enterprise). MySQL and Firebird will both need to be carefully watched in the future.

CNet reported on a mysterious slide that pointed out Sun's lack of a Sun built Database product and hint that it may be plugged with "SunDB". Uhhhhhhhh, this is a bit of a shocker. So, I'm jumping the gun, I know, but do we really need a database from Sun? I'm a zealot, no one will or can deny, but even I'm not so sure about this. This is problematic if for no other reason that the majority of midrange and highend systems that Sun sells are specifically to drive databases like Oracle and Sybase ASE. Oracle is really where its at, because Oracle provides more power and flexability than any other RDBMS on the market, open source or otherwise. Furthermore, when your shellin' out $40,000 per CPU for Oracle (or $35,000 for Sybase ASE) your not really going to be upset at the cost of high power midrange and high end SMP boxes becasue the cost of the hardware is now a shadow of the software cost. Example: A 12 CPU (UltraSparcIV Cu 1.2Ghz) E2900 with 96GB of RAM plus the trimmings only costs you $332,000, but Oracle Enterprise Edition runs you $480,000 (plus support at 22%, coming out at a whopping $105,000 per year!). So in a sense these seriously high cost databases help out Sun sales because suddenly it seems stupid to run your business critical database on a pile of Dell's or HP systems when you can have a highly fault tollerant and effecient UltraSparcIV based solution thats not a big cost beside the software.

Does Sun wanna push it's own database? I don't see why they would. If I were Jonathan and I just absolutely felt like I needed a low-cost/no-cost database to put into the growing Sun middleware stack, I'd not write my own... I'd put massive effort behind PostgreSQL. PostgreSQL is the best database avalible when you take Oracle out of the picture. DB2 is a lumbering dinosour, Sybase ASE is just bad value for money and still smacks of all the "Microsoft fscked us!" bitterness, and Informix..... please. So then you've got Firebird (pass), MySQL and PostgreSQL. MySQL definately has a major force behind it but just doesn't have the features of PostgreSQL, and won't anytime soon. However, we're still seeing a MySQL vs PostgreSQL showdown in the much needed area of replication and clustering. Both have it, but you can debate the good vs the bad all day long between the two. Right now the crowds are siding behinds MySQL replication but PostgreSQL is really really coming along [PDF]. Sun's always been fond of PostgreSQL anyway. If I could tell Sun to put devs on anything it would definately be PostgreSQL. We'll see what comes of all this.

UPDATE!: the story just broke on /.. Several people are pointing to Clustra as the possible SunDB. the acquisition press report makes it seem possible. Time to learn about Clustra.

The NetBSD 2004 Annual Meeting Report is avalible. Christos Zoulas, the president of The NetBSD Foundation, slammed just about all the major players while describing NetBSD's relevancy. And here is where I'm supposed to tell Christos that he's dead wrong and NetBSD is nothing more than a historical footnote, right? Be supprised if you like, but no! Christos has pretty well got it right.

Reguarding Solaris he said "Sun is trying to keep Solaris relevant by open-sourcing it, but nobody is certain of what is going to be open-sourced and when." Sure, I could give you a long long list of reasons that Solaris is relevant that have nothing to do with OpenSolaris, but I'll resist (Zones, DTrace, RBAC, ...). We learned about what is going to be released and when during the OpenSolaris preview, but I can understand some amount of confusion without having seen it with his own eyes yet. We'd like to do that, just put it out there, but we're not, as Scott says, "dumping code over the wall". We want this solid and ready to build on as smoothly as possible and that means putting some time into perfecting and modifying the build systems for external developers and getting the infistructure in place. Remember that Solaris has been under development for a long long time inside Sun and wasn't intended to be shared with the world so just geting everything ready for a newbe at home with his Athlon is a concern for us. We're shooting for quality.

Anyway, back to NetBSD. I for one have to admit that they present some good arguments to really keep excitement hot around NetBSD and to drive development forward even if they might be lagging a little behind the pack. Choice is a wonderful thing and NetBSD has a lot to offer, providing a really good choice for anyone needing a quality operating system. Even though I don't use NetBSD personally I hope that they stick around and thrive for a long long long time to come. Catch up on what NetBSD is up to on NetBSD.org.

Funny little article about Bill Gates' "interoperability" FUD. Hmmmm. Lets see, who is it claims to adhere to standards while ensuring that they completely mangle the spec at the same time? Hell, just look at Internet Explorers inability to handle W3C standards! Microsoft can try to bash open source all it wants but one of the reasons so many of us love Linux and frankly any non-Microsoft OS is the fact that the rest of the world plays by the standards. Linux, Solaris, AIX, HP-UX, IRIX, MacOS X, BeOS, QNX, etc, etc, etc.... we only have minor portability issues (typically endian) which any good developer will resolve whether they expect to port or not and standards (POSIX, W3C, OASIS, you name it) have provided us with the interoperability we need. So who isn't interoperable? It's not us. FUD, FUD, WTF FUD. Then again, when you've got that kinda cash in your pocket I guess your allowed to be somewhat delusional.

By popular demands we're going to take a whirl-wind tour of Solaris networking and configuration.

"I've almost just reinstalled Solaris to get the network settings fixed!" Ya, sometimes you can get alittle confused and irritated. Whether it's a new Linux distro, or diffrent UNIX OS, everyone puts their network configuration in diffrent places and learning each systems setup can be a pita. So, what if you don't even wanna bother learning all the network configuration stuff and what files does what... what if you could just re-run the networking configuration and have it all sorted out? You can! Solaris has a nifty command that can reset the base networking and configuration to a clean system: "sys-unconfig". If you buy a pre-installed system from Sun or a VAR, what they will generally do is install Solaris, customize the install as needbe and then run "sys-unconfig" before putting the system in the box and shipping it out. The command will scratch the networking configuration (most of it anyway), the root password, time zone, etc. Once sys-unconfig finished cleaning out the files it will immediately halt the system (shut it down). Next time you boot the system it will prompt you for all the networking information, time zone, locale and a new root password. This is the "quick-no-hassle" reconfiguration method. Generally we do not recommend this as ideal, but it will do the trick. If you plan on selling a pre-installed system or you are moving your system to a new network where all the info is changing, this is the command to use. Just use it carefully.

"I know root shouldn't login remotely, but I really want too!" Logging into a system via telnet, ftp, or ssh is just bad... very bad. However, some times people think they need to do this or they simply don't care about security. If you really wanna drop your pants like this, edit: /etc/default/login. In this file, you'll see a line for CONSOLE. To allow root to log in via telnet or ftp just comment out this line and your done. To allow root login via SSH you need to edit a diffrent file: /etc/ssh/sshd_config. Change the value of PermitRootLogin from "no" to "yes" and restart SSH via "svcadm restart ssh". The SSH configuration file is just the same as Linux since Solaris uses OpenSSH. Please note, that allowing root to log in remotely is a massive security risk and needlessly stupid, don't do this unless you absolutely must.

"How do I set and make the default gateway persistant across reboots?" The default gateway is easy to configure, just put the IP address of the default gateway in /etc/defaultrouter. If this file doesn't exist just create it and add the IP. The IP address of the gateway should be the ONLY thing in this file, nothing else. Typically you should just use echo, like this: "echo 10.0.0.255 > /etc/defaultrouter". Once you've changed the file you can restart the physical network interfaces to make the changes effective without a reboot ("svcadm restart physical").

How do I set or change my IP address? Setting and changing IP addresses on Solaris is a snap. Just add the IP and hostname that you wish to use to /etc/hosts. Then put the hostname (by itself) in teh /etc/hostname.(interface) file. Then just restart the physical networking via SMF ("svcadm restart phsyical") and your done! Example: "echo '10.0.0.25 solarisrox' >> /etc/hosts; echo 'solarisrox' > /etc/hostname.hme0" assuming that hme0 was the interface you wanted to use. Don't worry about this changing the hostname for the whole workstation, the hostname for the system is kept in /etc/nodename. You can add the default gateway as we discussed just above. If you need to specify a non-standard subnet netmask you can add it to /etc/netmasks. You can have as many subnet masks in the /etc/netmasks file as you like, one per line, just follow the instructions at the top of the file.

How do I configure DHCP? DHCP is pretty easy to configure, just create 2 empty files: /etc/hostname.(interface) and /etc/dhcp.(interface). If you are using HME0 you could do this: "echo "" > /etc/hostname.hme0 > /etc/dhcp.hme0". You might want to wipe out information that might conflict like the /etc/defaultrouter file if you created one. Once you've got these two empty files ready just restart the physical network to make the changes effective without a reboot ("svcadm restart physical"). If you want to customize the properties of the DHCP agent you can check out the " System Administration Guide: IP Services" manual. Remember that DHCP settings take acouple seconds to be retrieved from the server so if you reset the physical interfaces as mentioned above and you notice that nothing happened, just wait a couple seconds (30 secs is the default timeout).

I can't figure out what the network interface is called! Linux sort of makes network interface naming easy because typically interfaces are named "eth0", "eth1", so on and so forth. In Solaris, however, the name of the interface is decided by the interface driver. For instance, the onboard ethernet of a Sun Ultra workstation uses the HME (Happy Meal Ethernet, also more boringly called Hundred Meg Ethernet) driver and so the first instance is "hme0". You can look at the device mappings to driver names using "prtconf -D". In general, Solaris will find your network device and try to configure it so generally just checking "ifconfig -a" is enough to find it.

My l337 workstations network device isn't frickin' support!?! Sadly there are lots of networking devices that aren't officially supported (right now anyway, OpenSolaris hopes to change this). Thankfully there are tons of really kool Solaris developers that have been creating open source drivers! Check out these free NIC drivers by Masayuki Murayama! Grok google for drivers, you just might find your l337 gear really does have drivers avalible, they just aren't in Solaris right now.

There ya go.. some quick Solaris networking tips. As always, check out docs.sun.com's Solaris10 manuals for help and details.

A while back I wrote Eplayer (origonally called eVorbisPlayer), a small audio player that brought together the elegance and power of both the Xiph audio libraries and the Enlightenment Foundations libraries, using libao, libogg, libvorbis, Edje, etc. As you can see from the Eplayer interface, the buttons are oddly large... large, like the size of, perhaps, your finger! I created the interface with the intent to use it on Linux based PDAs such as the iPad and Zaurus. I went so far as to buy a Zaurus SL-5500, load it with OpenZaurus, and start cross compiling software for it. But, my dream fell short. Cross compiling can be a massive pain and there were several diffrent toolchains to choose from, all of which seemed somehow overly complicated. Eplayer's life changed dramatically when Trill started contributing to Eplayer, rewritting large chunks of it and massively improving it beyond what I'd origonally intended (I wrote it really simple and light to provide for both a neat app but also as a good learning tool for new E coders). Eventually we were contacted by the XMMS2 guys to see if we'd be interested in building interfaces for it, which was great for us because while we were having fun writting the interface we were having to spend more time writting the backend codec support, playlist handling, etc, all of which we didn't really want to do. They needed an interface, we wanted a backend, match made in heaven. The new version named Euphoria, Trill ripped out the backend to use XMMS2 instead of our own backend plugin structure and I at that point took my name off Euphoria leaving it completely in his hands since I wasn't really contributing anymore.

Ok, ok, long story I know... point is that I really wanted to transition my "desktop" app over to my Zaurus, but with all the problems I had messing around with cross-compilers and toolchains I just gave up in frustration, leaving my Zaurus little more than a nifty organizer (and I don't use organizers). But, last night I checked to see if any new images were avablible and learned about OpenEmbedded's BitBake. Wow! BitBake rules! It's a cross-compiling metadistribution... think of it as Portage for ARM but running on X86. Took me alittle bit to figure out how to fully set it up, but once I had things in place I was blown away! I told it to "bake" (build) EVAS, it then built its own toolchain, calculated all the dependancies for EVAS, then started downloading those, and even grabbed EVAS from CVS to do the build! Talk about easy and kool! I'm so excited I'm seriously thinking of picking Eplayer back up and primed for Zaurus by adding it to the BitBake repository.

So the moral is this: If you wanted to develop for Linux on ARM (Zaurus, iPaq, etc) but thought that it just was too big a pain in your butt to even bother, now is the time to try again with BitBake. Truely amazing work by all the guys over at OpenEmbedded!

James Dickens, my main man, put up some more kool Solaris tips! Find more tips from John Gardner on Solaris X86 printing and getting KDE installed, and some excellent pointers on using SMF by Liane Praza! This is kool, more more more tips guys! Lots of new folks, lets help them all out with all the tips we can cough up.

For starters, if you'd like to try out a real window manager, check out my Enlightenment DR16.7.1 packages! There are 4 of them, download each, install, update the runtime linker paths via crle, then reset dtlogin (the login screen, "/etc/init.d/dtlogin reset"), and off you go. Enlightenment isn't for everyone, but if your a hardcore ol' school geek at heart who doesn't like crap all over you desktop but you also want a huge feature set with pretty graphics and lots of eyecandy, then E is for you! Find screenshots of E in action on the Enlightenment website.

Speaking of packages... how do you install them? Easy, you use the "pkgadd" command. Say you download a package from SunFreeware.com or my Enlightenment packages, you'll generally download a compressed package. *.gz is GNU Zip (use "gunzip"), *.bz2 is BZip2 (use "bunzip2"), and *.Z is Compress (use "uncompress"). Then use the "pkgadd" command, like this: pkgadd -d ./CUDLsomething.pkg. The -d points to the package, but watch out because you must prefix the pkg with a path, relative or absolute! If you say "pkgadd -d CUDLsomething.pkg" the command will fail saying that it wasn't found, but "pkgadd -d ./CUDLsomething.pkg" will work fine. The reasons for this are more than most people would care about. Please note however that you must add each package one by one! Again, the reasons for this are more than you probly want to hear, namely that there are 2 ways to create packages, one is as a directory structure where you can easily add multiple packages at a time or alternately as a single file which is more portable but not as easy to install.

To manage or remove packages you can use "pkginfo", which will by default list every package on the system. If you want to learn more about a package, such as SUNWogg-vorbis, you can get the long pkginfo output via "pkginfo -l SUNWogg-vorbis". If you wanted to remove this package, you could use pkgrm, such as "pkgrm SUNWogg-vorbis". With pkgrm you actually can remove multiple packages at one time, just list one after another. If your interested in learning more about Solaris Packaging take a look at the Application Packaging Developer's Guide.

But what about patches! You can look at patches in several ways, but the most straight forward is to use the command "showrev -p". The "showrev" command can show you all sorts of information but with the -p flag it outputs a list of all installed patches. You can download patches on Sun's support site SunSolve. On SunSolve you can download patches individually, in a recommended patch cluster (download 'em all and install with one command), sign up for the "Patch Club" or request notifications for critical or security patches, and you can also download (all for free) diffrent patching tools that can automate and simplify the proccess of checking your current patch set, download the new patches, and install them. Using tools like "Patch Manager" can really simplify your life because some patches have various dependancies and restrictions which have to figured out one by one normally, but are all handled seemlessly by Patch Manager. Even better than that, Patch Manager is included in Solariis10 out of the box (well, ISO for now)! Want to check for any new patches that you need? Just run "smpatch analyze" (smpatch is the Patch Manager binary) and it will check for patches and tell you what needs to be installed, then you can use smpatch again to download and install them. Easy as can be. Check out the smpatch(1m) man page for all the details on what Patch Manager can do for you. You can read more about patching in the Solaris10 Basic System Administration Guide.

Can't get DNS configured? Did you forget to enable DNS during the installation and now that it's running your screwed? Don't worry, you need to edit "/etc/nsswitch.conf". There are two easy ways to do this: 1) edit /etc/nsswitch.conf and add "dns" to the "hosts" line, or 2) Even easier is to just use the template by copying "/etc/nsswitch.dns" to "/etc/nsswitch.conf". Once you've got nsswitch.conf set properly just go ahead and edit /etc/resolv.conf like you would on any other system and DNS should work. The nsswitch.conf file on Solaris tells each name service where to look for information. By default all the name services look in "files" (the normal files in /etc), but you could also specify that Solaris should check nis, or nis+, or dns, or ldap in this file.

Lookin' to export NFS filesystems? There isn't an /etc/exports, so you might wonder just where the thing is... You want to edit /etc/dfs/dfstab. In the header of the file are instructures on how to get things working. To export filesystems we use the "share" command instead of "exportfs" in most cases although the "exportfs" command is avaible on Solaris just like your used too but all it really does is translates normal exportfs arguments into "share"/"unshare" arguments. If you couldn't guess, "share" will share a filesystem via NFS, "unshare" unshares it.

More tips to come. Everyone out there blogging Solaris, add your own tips for all our new friends. We all had questions once, lets save them the trouble!

Need a good laugh? I've never bought into the idea that blogs would replace traditional news outlets, untill I saw this: Sun, Open Source and Star Trek. At this point I just can't refrain as usual.... what a fscking idiot! [edited]! Think for yourself for once and don't just buy into what you think people want to hear. Here's my favorite frickin' quote:

Listen, Simon, Scott, Jonathan, forget about calling what you're doing open-source. No one says you have to be open-source. Well, OK, so you need to do something about the fact that Linux is eating Solaris' market share lunch, but I think you've already got that with your N1 grid model.

Yes, someone is saying that Sun needs to open Solaris... its community of developers and administrators for the last several years. Get off your high-horse and drink some water, maybe even breath in and out alittle bit. Whether you personally get it or not Steven this thing is happening... its been happening. Look through my blog a bit and look at the proof for yourself. This is just a huge pile of FUD from old school morons who didn't get Linux in the mid 90s and don't get OpenSolaris now. Holy shit, have you stopped to think that when members of the community and customers talk to Sun one of the things we keep saying is to resist the urge to cater to this kind of FUD crap you guys are spewing? Stay the course, do it right, take your time, evolve the industry, thats what we're telling Sun and thats what Sun is doing. Spend alittle less time around elitists and more time around real developers please and perhaps read some history over the weekend. If you still don't belive it, maybe you should get your hands dirty and join the pilot to see for yourself. Please stop and THINK for a second that there are more than 60 non-Sun pilot members who are pouring over code and trying to work, while jackasses tell us that this isn't open source and that we don't exists, and that this is a huge mistake, and that we're all evil, blah blah. Pure, stupid, FUD. You guys bashed Linux and now your bashing us, so it's hard to put any stock in the endless pile of crap you guys dish out. I'm so sick of this shit. Sun has to put up with your bullshit Steve but I don't.

With all the people now downloading and checking out Solaris10, I thought that maybe alittle quick primer might be in order. Here are some of my personal favorite features in Solaris10 (some exist in earlier releases as well), specifically if your coming from Linux or having played with Solaris in a long time this might be of help.

Filesystem Layout: How to find stuff: When I think of the way SystemV should look Solaris just seems to be the model. Some basic overview is:

• /etc, /proc, /devices, /dev: Samo samo, you know what these are
• /export: More common on Solaris than Linux, typically this is where "everything else" goes. Some people even put local (non-automounted) home directories here.
• /opt: Applications, just like Linux this is where you'd typically find 3rd Party installed applications
• /usr/bin, /usr/sbin: Samo samo, user and admin apps
• /usr/sfw: You wanna look here! This is where the Sun FreeWare applications (GCC, GNU Tar, etc) tools are, including libs, headers, etc.
• /usr/ucb: UCB standing for University of California Berkeley, these are the BSD version of tools, mostly for compatablity. If you prefer the BSD version of ps (ie: ps aux) you can find it here.
• /kernel: This is where the modules and their associated configuration files are.
• /platform: This is where platform specific code is held, namely your kernel (the kernel is named "unix", similar to the Linux "vmunix")
• /cdrom: Automount point used by the volume management daemon (vold)
• /home: By default /home is an automount point! If you want to use it locally first comment out /home in /etc/auto_master and then restart the automounter ("svcadm restart autofs")

Finding out what you got!: You can determine info about your system using "prtconf" (outputs the device tree, driver attachment information, etc) and "prtdiag" (found in /usr/platform/(platform)/sbin). Wanna check how much memory you have, or CPU, use these.

System Statistics: Most Solaris sysadmins are content with vmstat and iostat for information, we then extend that with truss, lockstat, etc. In Solaris10 we can now go further with DTrace! (Previous to DTrace alot of used the SE Toolkit) You might be wondering where "top" is though, the Linux favorite... we don't have it! Instead, Solaris has it's own version: "prstat". We have lots of other ptools that can give us information based on the /proc filesystem (you'll notice that /proc on Solaris doesn't have all the general data provided by Solaris, we get it in diffrent places like "prtconf"). We've got pgrep, pmap, pkill, pargs, and more. For info on the ptools check the "proc" manpage.

SMF: Ding dong the witch is dead!! The Service Management Facility is the first real replacement of the aging RC Init system used by everyone else. No more doppy scripts! SMF allows services to have dependancies to other service, so for instance if you stop your ethernet device it can also stop all the networking services dependant on it too, and stuff like that. "svcs" can report Service Status, "svcadm" can control services (enable (think start), disable (think stop), restart, put in maintance, etc), "svcprop" can manage service properties, and "svccfg" can configure services. Start out by just running "svcs" and looking at the list. Then try using "svcadm" to restart acouple things or enable and disable services.

Apache2: Solaris ships with alot of goodies, including Apache2! Just edit the configuration (rename from sample) in /etc/apache2/, use svcadm to enable the service, and then drop your pages in the /var/apache2/htdocs directory! Apache2 provides a great test for Zones.

Dev Goodies: You can find the bulk of the standard dev tools (as, ld, ld, lex, make, etc) in /usr/ccs/bin. You can interface with the runtime linker by using the "crle" command (Configure Runtime Linking Environment), rather than edit a file like on Linux.

This is just the begining! I won't bother posting more right now, but you can find a ton of information in the docs found at docs.sun.com. Find information about Kernal Tunables, Security Services (IPfiler and IPsec goodies inside!), Solaris's included Volume Manger, Zones and Resource Management, a complete set of development docs, and of course the DTrace users guide. Solaris is chop full of goodies, you just need to know how to use them. More goodies another time.

I've been checking Oracle.com just about every day waiting for official Solaris10 support for Oracle10g. I've been running 10g on Solaris10 for several months now actually, and it runs great, but I've had all sorts of strange troubles trying to run Oracle10g RAC in Zones. I thought it'd be handy for testing purposes and then saw someone trying to do it in the Sun Forums and thought I'd help him out... but hit my own problems. I'm hoping that this isn't just me and is actually a compatability issue, but maybe it is just me. Either way, I'm on the lookout.

Anywayyyyyyyy, the point is that on the Oracle News page I found this: Multi-core Processors - Impact on Oracle Processor Licensing. Long story short, Oracle considers a "processor" to be a core. This is diffrent than other software vendors who have moved to a "per socket" pricing model. Oracle says that no change is coming, although they are considering per-employee based pricing to help control costs for customers, which would be welcome. Current standard Proccessor Perpetual (all-you-can-eat based on CPU(core) count) is $40,000 for Oracle Enterprise Edition, what most people run. This means that on an [PDF} 8 core Niagra could cost you $320,000 for a single socket system. Ouch! Oracle Standard Edition ONE however is a more reasonable $5,000 Proccessor Perpectual making a Niagra implementation only $45,000 (plus support at 22% of the purchase). Apparently IBM is doing the same thing with DB2 and BEA plans to charge a 25% premium for multi-core procs. I looked for word on what Sybase is doing with Sysbase ASE, but didn't see any info. Not that it matters, the progress made by Oracle is so significant that I don't see any reason that someone would seriously consider DB2 or ASE on any basis other than cost. MySQL and (my personal fav) PostgreSQL unquestionably have their places in terms of functionality and cost, but when comes to outright power and flexability you just can't beat Oracle10g. DB2 is in typical IBM fashion a mess and ASE just feels kinda flimsy. We'll see where this core-vs-socket argument ends up in another year or two.

The presentation hasn't happened yet but the slides are online![PDF]. Check them out, the presentatin is called "Communities Matter". Go to to listen to all the events associated with the Sun Analyst Summit going on yesterday and today.

See, this is the stuff I don't like.... politics, legal crap, etc. It's time spent dealing with things that don't directly advance a project from the standpoint of the individual developer, hence the desire to just use MIT or BSD licenseing, or go with the GPL flow, and be done with the whole thing. Figured I'd put up some links to various things on the topic of interest:

• SGI open source Projects
• IBM open source Projects
• IBM's Common Public License (CPL) and FAQ
• The IBM Public License
• IBM's much talked about 500 free patents [PDF]
• Sun open source Projects
• Sun's Common Development And Distribution License (CDDL) and FAQ
• OpenSolaris web site
• HP open source Projects

Maybe that helps people out. I couldn't find any projects being contributed by Dell, more than likely their just working closely with Red Hat rather than doing things in house. If you look at the sites above, SGI has really contributed quite a lot and was one of the first to get serious (XFS rules!). IBM and Sun are both major contributers. I find it hard to say that one is more commited or contributes more, they simple differ in what they contribute too. For instance, both IBM and Sun have contributed to Apache and Mozilla, but IBM tends to really like some deeper technologies closely linked with Linux such as Xen while Sun focuses more on desktop and infrastructure project.

People keep trying to duke it out with me over whether Sun or IBM is more commited to open source. And really, it's a hard discussion to have because it depends on which side of the street you are and where you want to go. IBM generally seems to embrase open source projects as their own, as an example look at this: sHype Hypervisor Security Architecture for Xen. In this one example IBM didn't march in and take things over, or fork, or anything, they decided that Xen was something they liked but needed to extend capability further than it current supported and they'd leverage existing R&D (sHype) to do it. Thats definately a sensable and really kool way to approach it. IBM's GPL offerings have been nifty too (JFS rules too!). In general you get a feeling that IBM is leveraging existing project to move their offerings in the direction they want by combining IBM R&D and engineering with community efforts in a pretty smooth way. "The communties got it, we could use it, we can help", might be what IBM could say.

Sun takes a diffrent angle. Whereas IBM is leveraging a lot of external open source projects and basing offerings on Linux, Sun isn't looking outward, but inward. Sun (and it's userbase) like the technology already avalible but want to extend and build on that rather than simply embrace Linux like other vendors are doing. Case in point, Sun could have embraced Abiword or Koffice, but instead chose to start up OpenOffice. Both IBM and Sun had a choice, in a sense, open up your own OS (AIX and Solaris) or just embrace whats already open (Linux). IBM choose to go with the external code (Linux) and Sun with the internal code (Solaris). There is nothing inherently evil about this. Much of the anger seems to be that Sun was supposed to (so they think) join the Linux tea party and leave Solaris behind like IBM choose to, but thats just not in Sun's blood. Sun, mainly due to Scott's firm leadership (say what you will), charts a pretty straight course, making small divergances here and there to support something new, but not major shifts. IBM is, in a sense, used to supporting multiple OS's, looking just at z/OS, AIX, OS/400, and now Linux to name just a few. Sun doesn't do that, they won't split dev into two equal Linux and Solaris groups, because Solaris is the Sun direction, Linux is a Sun choice (want it? buy it from Sun) but not a source of major development. But the open source world is much much bigger than just kernels, which is why Sun is involved in alot of non-kernel projects.

Sun is giving to the world what it has to offer and IBM is giving to the worlds offers what it has to add. These are two polar ways to go about it all, and naturally why so many battles break own. If your on the Linux side of the street your naturally gonna prefer that IBM add features and functionality to the Linux kernel. If your Solaris side of the street you really feel like your missing out on the ability to extend and modify that platform for your needs like the Linux guys can extend and modify Linux. Neither side is wrong, but Sun believes in it's products so much that it will open them to the world and let everyone play rather than just toss in the towel and move to Linux, whether you agree or not that should mean something.

In October I got re-engaged with USENIX and SAGE, the System Administrators Guild. Both are very good organizations seeking to assist and provide an organization to unify sysadmins. Of course, with any organization that requires dues you eventually ask yourself "what do they do for me?" Having USENIX/SAGE on your resume is always a good thing in and of itself, but is it anything more than just that and some discounts? I think so. I've always been bummed out that as a lowly and poor sysadmin I've never been able to afford a ticket to the various confrences they put on such as USENIX, LISA, or even FAST. Despite this the organizations provide value in connecting us as a profession and working on our mutual behalf.

One of the ways that SAGE has attempted to accomplish this goal is to provide a sysadmin news site: SAGEwire. Sadly, the site feel into neglect. I approached SAGE in October about getting it back into at least a basically managed state, or at the least let me have admin/editor access so I can at least post stuff. They agreed and I've been trying to keep things up to date. Sadly, because I'm also a contributer to Bill's excellent SunHelp.org, and I run things on cuddletech, my own site, I sort of just repeat myself in 3 places when posting news or end up neglecting one of them. Because I spend most of my time on Solaris these days I don't get to submit as much HP, DEC, SGI, and IBM news (technical, there is no need for he said-she said on SAGEwire) as I'd like to have. Basically, I need people to make submissions to SAGEwire. If you think the site might have value, consider submitting stories once in awhile... if not, well, leave it and let it rot into oblivion. I'd like the site to live, I think its a good resource, but one person can't sustain a good news site for long, especially when the breadth of news that the site requires is beyond what any one person can provide. If it sounds interesting, check out SAGEwire and consider submitting from time to time, at this point I'm accepting any non-advertising submissions. [FYI: My name isn't on the editor list even though I'm approved, therefore I submit all stories as "sage-admin". Anything you see from "sage-admin" is me.]

Thought I'd toss out some little nuggets sitting my array of tabs... I just learned of Brad Marshall's Blog, very kool guy that has been really great to chat with in #opensolaris. Sid Art made a kool post to his blog entitiled "The Linux jihad against SUN", I'm not sure I'd agree with every word but damnit he's got style. ZDNet is coming around with a nice, and oh so refreshing, article entitled "Sun gives way more than an inch, but open source advocates want the whole mile" by David Berlind, bless you David, bless you. Over at Slashdot, someone pointed me to pkgsrc: The NetBSD Packages Collection, which looks to be nifty, I'm still playing with it. Also, in digging around the net I came across this paper "Software Distribution Scenarios[PDF] by Thomas W. Storey, it's got a pretty nice overview of the popular open source licenses in it which is probly a good "if you only read one thing ..." overview in it. And lastly, as seen on Slashdot the other day the US Army Basic Cryptanalysis Field Manual, which is a great and facinating read for any of us crypto enthusiasts.

We put Nova to bed about 9pm as usual and Tam decided that she really wanted a milkshake, something you just can't say no to. So I set out into the night and picked up some goodies on the way. I hit Target to nab the AVP Widescreen DVD, because the movie just ruled. If you thought AVP was lame you either don't like SciFi, or your the kind of person that thinks most adult films don't have a well developed plot. Since I was out, I ran by Borders to pick up some books. A perk of my job is that I get re-embursed for books, so I thought with all the license talk I'd brush up on things, so I nabbed: "The Cathedral & The Bazaar", "Userstanding Open Source & Free Software Licensing", "Write Great Code", "Version Control with Subversion", and "Linux Kernel Development. I probly shouldn't admit that I don't know Eric's book but I've never felt that it was important for me to read, and as for the licensing book, I've always felt that the license should speak for itself without the need to be explained to you bit by bit, but now that I'm discussing licenses more actively I need all the reference I can get. I've wanted to grab the "Write Great Code" book for awhile cause it looks pretty sweet, and the SVN book I've used alot online but figured I'd shell out some bucks for the authors and keep a hardcopy around. As for "Linux Kernel Development", Robert Love is just a kool guy and a great writter, so I thought it'd be a good read, and hopefully it's a good informantive read because I've not really liked any of the other "Become a Kernel Hacker!!!" books so far... I freely admit that I'm not a kernel hacker and I can't write device drivers, but I really would like to at some point. Dropping out of college can have its downpoints sometimes.

Anyway, I have a point... I finally pulled into the drive through of In-And-Out (Californian hamburger joint) for my shakes and end up having to wait 30 frickin' minutes in the drive through, and you can't leave the line. So I pull out the copy of "Linux Magazine" (Feb 2005) that I picked up (nice artical on M4 in there, hard topic to write about) and started flipping through. Then I found the last page... ouch. The closing page was written by Jason Perlow and he put out his feelings about the CDDL and OpenSolaris. Now, bear in mind that he wrote this just before Christmas so he was making the guess (that a monkey could have made) that CDDL was in fact intended for OpenSolaris. So, in this little rant, he gets acouple things really really wrong which I wanted to address...

For starters, he complains about what many others are, GPL incompatability. But, like I've said before in this blog, it's not the CDDL thats incompatable with the GPL, but the GPL thats incompatable with the CDDL. CDDL provides for sublicensing, GPL does not. MPL isn't GPL compatable, which forced them to turn Mozilla into a triple threat, relicensed under MPL/GPL/LGPL allowing the contributer to choose the license they want. This insistance that GPL is the end all be all of open source licenses is somewhat silly. There are lots of good licenses avalible because there are lots of good ideas on how licensing should work, and if you write it you get to choose how its licensed, whether its me, or you, or Sun.

Then he suggests that there should be a GPL Solaris and a proprietary "value added" version which would include Zones, DTrace, PSH, etc. Ya, sure, like releasing Solaris10 as open source with all the goodies removed is gonna make people happy.

Then he slams Sun for putting GPL apps in Solaris but not using the GPL itself. Not fully true because Sun contributes a lot of code to GNOME, a GPL project, and several others. He suggests that Sun's strategy "is a one-way street", but apparently he's just not paid much attention.

Then he goes on about how great BlastWave is, which is a good thing, although he calls Solaris10 an "otherwise spartan" OS. Of course he's A) commenting on Solaris Express, not Solaris10 GA, and B) he apparently didn't find /usr/sfw/bin where all the nifty goodies are. Solaris10 includes JDS 3, so out of the box you've got a GNOME (JDS) desktop with GAIM, GIMP, Mozilla, Evolution, etc, etc, etc, apps ready to rock, not to mention several GNU tools including GCC 3.4.3.

FInally, he bitches about Sun not supporting Blastwave.org. As he says "That's right, you heard me: Sun won't even support its own die-hard Solaris fans." Ooops, wrong. No fact checking I see. It isn't my place to disclose what Sun did or didn't do... ask Dennis about it if you like, but I can tell you that Sun was (and is) very interested in supporting Blastwave and Jonathan himself has been in contact with Dennis many times. As a, so called, "die-hard Solaris fan" (I prefer zealot) I can tell you first hand that Sun is not affraid to help out those who need it, provided its appropriate. There are plenty of times those of us in the community have argued about what Sun should or shouldn't support, but they've certainly never been afraid to engage in the discussion and do what is right for everyone concerned. Sun can not, and will not, get in the business of handing out gear just because someone is commited, it sets a bad president. We don't need a community of people doing good things just so they might get a handout from Sun, or a job, or whatever, we need people commited because they simply want to be commited to something really really kool. If support comes from that relationship, so be it, but neither Sun nor the community need people walking around with their hands out. Sure we'd all like free gear, but thats just not what this is all about.

So, if you pick up Linux Magazine this month (check out that m4 article!) just do yourself a favor and skip the back page. I'm a little suprised the editors didn't pull it.

Sidenote: I forgot to mention, apparently Jason was planning to ride the Solaris train with MySolaris.org which he started in August, but it looks like it fell silent in late Nov. Not sure if that means he was so bothered by the CDDL that he gave up or if he simply didn't have time. Interesting site anyway.

So we've been getting reactions from people slowly, one by one. Today at the OSDL Enterprise Linux Summit(tm) right here in the Silicon Valley Linus shared some thoughts on OpenSolaris. I was hoping to go to this event but had to perform a system upgrade that I've waited a year to get and couldn't pass up (gotta love Solaris uptime). Linus was, in a word: Linus. You gotta love him, he's just so mellow and layed back about thing in general. The real take away was: "It all looks good. I was disappointed in their Java work, it was a complete disaster, and Sun took control of it. But CDDL is different. Everything is in place for it to work well." This is the most senesable and welcome response I've heard yet, and I'm really happy that Linus has this view. Of course, another thing I value about Linus is his candor, because I'm the same way, so I just loved this comment: "A lot of people still like Solaris, but I'm in active competition with them, and so I hope they die." You gotta love this guy.

In the same article, we also see mention of Danese Cooper, the so called "Open Source Diva" of Sun, saying "We're not going to sue anyone." She's "Senior Manager, Open Source Programs Office", so if you wanna knock someones door down, she'd be the one I guess.

Don't forget that Sun's Network Computing 2005 Q1 quarterly announcement is today at 2pm PST. Sadly, I've got a schedualed maintance happening this afternoon at 1pm untill whenever, so this is the first NC event so far that I won't be able to watch live. While your hearing about the Sun Grid and new gear I'm gonna be installing patches and fixing some issues on my E420R in a chilly server room (and after waiting a year to get the outage I can't put it off).

In other news, ZDNet has an interesting article entitled "Sun: Patent use OK beyond Solaris project", hopefully that will set some people at ease, but we'll see what the fallout is. As usual, in the eyes of some people out there Sun can do no right these days. For more goodies, listen to David Berlind's "IT Matters" podcast. In todays podCast he talks with Jonathan about the announcements. Just skip the first 30 minutes of it, it's all about the Grid stuff ($1/CPUhr), at about 35 minutes into it they change gears to talk about OpenSolaris and CDDL and patents. Just listen to that last 15 minues or so.

Lets hope for some good press outcome post NC today... I've got my fingers crossed, you just never know how its gonna go. Half this stuff just goes in one ear and out the other with the press, but lets wait and see what they do.

Ulrich Drepper of Red Hat has posted a [...] message on the Red Hat GlibC dev mailing list, subject line: "regarding the Solaris trap". I won't repost it here, read it for yourself. So much for OSI approval meaning something eh? I'm just alittle supprised that people are looking at OpenSolaris as nothing more than a Pick-n-Pull fest that suddenly got rained on by the CDDL. Is GNU code so bad that it needs to gut any goodies in the OpenSolaris codebase? Of course not! So many people just seem to be approaching this whole thing in the wrong way... CDDL wasn't expressly "intended to be incompatible with the GPL", it was intended to guard some freedoms that the GPL doesn't provide. GlibC is of course LGPL, which is imho a much nicer license. (FYI: in some cases I personally use the GPL specifically because it's viral, but I don't think its for everyone and everything. In the future I will be considering CDDL instead which is a much nicer balance than GPL vs BSD.) Honestly, did people really think that the minute OpenSolaris code was avalible it'd be ripped to shreds and integrated into other projects? Sun isn't out to sue the world, and if they ever looked to be "dubious" I'd be pounding down the doors, I know where Jonathan's office is and I can buy a megaphone in a pinch. This isn't a trap! It's simply an open license that protects the developer, the contributer and the distributer and I for one think its a really well found balance. Sorry to rain on your parade, but CDDL provides a fair and level playing field for all sides involved, too bad if that means a rip-o-rama can't be put on the calander.

I can tell you one thing for certain. As a pilot member I saw and reviewed the license (along with the other pilot members, internal and external) prior to OSI submission. As soon as I got the license I printed it, got a cup of coffee and when out side to chain smoke for about an hour while marking it up. Some changes needed to be suggested and they were accepted by Sun. After reading over it acouple times I liked it but re-read it with the "Would I use this as a general license for my own code with no relation to Sun and/or OpenSolaris?" and the answer was yes. Several good options are around, from "I don't give a rats ass what you do, just leave my name on it" (BSD), to "Friends don't let friends not GPL, touch this code and you too will be GPL... because thats freedom" (GPL), to "You respect me, I'll respect you, share alike and prosper" (CDDL). Anyway, the point is, I'm not evil, I'm not dubious... I'm just a guy at home that likes to code and thinks Solaris is a really great platform even though I like Linux in some instances and I definately like the GNU but don't think it's for everyone... I'm not plotting against Linux or looking to make money or cash in, but I signed off on CDDL. Its the right thing, its the responsable thing. Why thats so outragious I don't know. Remember, it's not the CDDL that won't play with GPL, it's GPL that won't play with CDDL. CDDL and BSD can combine with no problems.

Stay tuned for NC05 Q1 later today... hopefully Jonathan addresses some of these concerns.

Finally Finally! Solaris10 GA can be downloaded now for your personal and professional enjoyment. Sun.com is getting pummled right now too, I'm getting only 60K down at the moment and it's just dropping like crazy.... and damn thats a good sign!

For anyone who hasn't run Solaris in awhile and is now trying it out, some words for you. Solaris on X86 (32bit and 64bit) is amazing [if you notice your AMD64 system boots a 32bit kernel from the CD, don't freak out, it'll automatically install 64bit support, it's only 32bit during the install], but driver support isn't nearly as good as Linux. No ones going to be shocked by that fact, but don't assume that Solaris/X86 supports everything that Linux does, because there are some holes. If you intend to do some basic hacking with it, your good to go, but if you expect hardware GL accelloration from your nVidia 6600GT, your going to be disappointed, if you have the latest and greatest SATA chipset, your going to be disappointed. Basically, any time you play with Solaris/X86 and think "damn! the driver support sucks ass!" just ask yourself if other non-Linux OS's have better support, such as FreeBSD or QNX. Linux is way ahead of the pack when it comes to driver support which is why many of us run it whether we think it's the best OS or not.

Some kool stuff to check out in Solaris10 is the SMF (Service Management Facility), part of the Self Healing goodies, which is Solaris10's replacement for the ancient RC Init scripts (/etc/init.d/). Look at the output of "svcs" to check out whats online (started) or offline (stopped) and play with "svcadm" to enable, disable, put in maintance, etc, diffrent services. Also, check out Solaris Zones (aka: N1 Grid Containers). And of course, your probly dying to check out DTrace.

Feel out the whole OS, but remember that on X86 Solaris is first and foremost a server OS, but looking at screenshots that I've posted you can clearly see that it's way more flexable than that. You might notice that while JDS might not be all of ours first choice in window management you get some kool apps ready to go out of the box. Don't forget to sneak a peak in /usr/sfw/bin!

If anyone thinks it'd be helpful for me to post some little mini-projects to help you get started in your l337 Solaris10 hacking ventures, just let me know! Maybe a new paper entitled The Cuddletech Guide to l337 Solaris10 Features is in order. :)

JDS/GNOME. KDE. WindowMaker. All just suck for real work. Things in your way, stuff to deal with. Only Enlightenment provides me with a truely restful enviroment to work in without clutter and hassle. Geographical virtual desktops, multiple desktops, superior stacking layers, ability to remember the window settings, easy window skip list management, ultra fast wallpaper selections, on and on. This is the way a desktop should look and feel...

OpenSolaris running on an Ultra2 [2x300Mhz UltraSPARC II, 512M] workstation, using Enlightenment DR16.7.1, Eterm 0.9, etc.