Tim Foster's Weblog

I presented An Introduction to OpenSolaris last Saturday at OSS BarCamp - my contribution to Software Freedom Day 2009.

You can download the odp presentation or the pdf, which I've exported with my notes for the talk that explain each of the slides a little more - I hope this is useful if you're planning on giving a similar talk.

A few things struck me while preparing for and giving the presentation. Firstly, it seemed odd to be giving an introduction to an operating system that's been around for quite a while now: clearly we haven't been doing enough of this sort of thing (and from personal experience, yes, ie-osug isn't as active as I'd like, I just sadly don't have the bandwidth)

Then secondly, it's really hard to cover all of the interesting features of OpenSolaris in sufficient detail over the course of an hour. My take, was to try to whet the appetite, rather than explain every feature fully - and in some cases, go for the features I thought the audience might be interested in (for example, mentioning NWAM as one of the major networking features - perhaps it's not as full of rocket science as other aspects of Solaris networking, but it makes a huge difference to the novice user)

Finally, and slightly embarrassingly, I had to spend about 5 minutes in front of an expectant audience futzing around with the display settings on my R500 laptop to get it to talk to the projector. It was doubly annoying that both xrandr (and indeed gnome-display-properties) were able to see the separate screen, but try as I might, I couldn't get any output to appear externally. Ultimately, a kind audience member offered me a USB key by which I transferred the pdf over to my EeePC (running nv_122) which was able to see the projector, but didn't have any of my demo material setup (some ZFS settings, some zones, crossbow, flows etc.) Oh well.

Here's hoping that at least some of the audience left with an impression that OpenSolaris was worth taking a second (or perhaps a first?) look at, despite the brevity of my talk and the initial teething problems I had. My new mantra:

Never work with children, animals, or weird exernal VGA projectors.

Some of the work I've done recently involved some changes to virt-install(1) to teach it how to install xVM guests from OpenSolaris AI servers - work that you'll see going back soon as a patch to virt-install as part of our xVM 3.3 changes.

This ended up shaking a few bugs out of AI and OpenSolaris, two of which became stoppers for the 2009.06 release (which made for a rather exciting weekend) one was fixed, the other was documented in the release notes.

Along the way though, and the point of this blog post, I got to learn a bit more about OpenSolaris and the boot process when we're using AI and what to do when things go wrong.

For x86 (the only thing I really cared about in my case, sparc differs slightly) AI works by downloading the kernel and a very small boot archive via pxe and tftp. The client then boots with this image and the svc:/system/filesystem/root:live-media SMF service arranges to download solaris.zlib and solarismisc.zlib files, and mounts them on the client.

However, should that service fail for some reason - we're left with a pretty unfriendly OpenSolaris environment. There's a tradeoff between fast/low memory installs and easy-to-debug environments so it's a tough one to call.

However, if you do need to debug stuff early in boot with an AI image, I added a comment to Defect 6851 that explains how you do it. This came from an email I was writing to a colleague today who was running into the same problem - I figured posting those comments to the bug report and writing this short blog post would be a good thing to do. Hope this helps someone out there?

It's with regret, that I won't be able to head over to CommunityOne West next week - the OpenSolaris tracks look pretty interesting, but as ever with this sort of thing, it's as much about the people you'll meet as the actual content of the sessions - in fact, I'd almost argue, it's more about the people you'll meet, than the sessions.

It's been ages since I've chatted to OpenSolaris-folk in person (other than my immediate colleagues in the xVM team and Glynn of course) and while email and IRC are good, they're no really match for having a beer and a natter with like-minded people - the OpenSolaris Party on Monday night looks like a pretty good opportunity for that. I would also quite like to be over there for the launch of 2009.06 having played a bit of a part in this release too (note to self: filing bugs that end up on a stopper list makes for somewhat exciting weekend), oh well.

My travel plans this year have been put on hold for a while - having been away from the wife + kids for a few weeks while I was down in New Zealand, closely followed by another week in MPK, I've promised to stay put in Ireland for a bit, although more about that in a future blog post I think.

Still, I'll be doing my best to keep up with what's going on at C1, and hope that I can at least read some blog reports on how the event goes next week. Wish I was there! If you wish you were there too, but live a bit closer, and feel like learning more about OpenSolaris, then do fill in the box below :-)

I voted:

RECORDED:  ballot 971674dbf07ccbd25a1bf7935a61ecc1a8b26493 on "Board
Election 2009/Change Constitution" from Tim Foster
Connection to poll.opensolaris.org closed.

- "yes" to what looks like an excellent change to the constitution, and for a set of seven people (I expressed preferences for all sixteen candidates) that I really would love to see on the OGB this year.

Yes, I voted before seeing much in the way of electioneering from several candidates, my view being that people going for posts on the OGB should be pretty well known in their work on the project already and I shouldn't only be hearing about them in the days running up to the election.

There's detail on the above on the 2009 elections page.

We've got the venue confirmed: here are details of the upcoming Irish OpenSolaris User Group meeting:

Topic	General OpenSolaris Discussions
Date	Thursday 29th January 2009
Time	19:30
Location	The Vaults

Look forward to seeing you there! If you can help out with equipment or have ideas for presentations, or just feel like saying "hi", drop mail to the mailing list.

I sent some mail to the Irish OpenSolaris User group list today, proposing to kick-start our user group meetings again.

Meetings haven't happened at ie-osug since last February, and we're trying to see if a change of tack would help get things going again.

Our meetings from June '06 - Feb '08 were more like a mini lecture-series about OpenSolaris, and while I think these were interesting, they often came across as a bit formal: yes we had pints afterwards (which were usually great) but there was never the atmosphere of community we'd hoped for.

So, this time, we're going to try the approach the SFOSUG use - try holding the meetings in a pub. The location we're thinking of, The Vaults serves food & beer and we'll hopefully be able to reserve a small room in the place. We'll bring along a wifi acccess point and a few laptops, and an LCD projector. We'll still be able to do "feature presentations" if people feel like doing them, but hopefully the more informal atmostphere of a pub will help get people talking a little more, and perhaps grow the user group and get more people participating.

I don't have the exact time/date yet - but will post more when I have it, I suspect it'll be Thursday January 29th. Do please comment here, or send mail to the list if you think this would be a good way to get more people interested in OpenSolaris in Ireland?

(though I'm happy that's being announced too)

Last night, close to midnight, I'd started writing a blog entry about Solaris on the desktop, the different environments we've used, and was going to talk a bit about how far we've come from when I joined Sun back in 1996 up to today's release of OpenSolaris 2008.11 (I was also going to mention my role in writing a chunk of Time Slider for the current desktop environment) but that post's for another day.

There was a much more important delivery today: Calum Henry Foster, 7lbs 2 oz, was born at 6:04am about 30 minutes after we arrived in the hospital. And yes, I broke a few red lights on the way in! Mother & baby are both doing well and I am the proudest father in the world (again!) - I'm grinning from ear to ear right now.

Yow - October already, how did that happen? I'm a bit late with August's OpenSolaris monthly news, and now September's has piled up on me as well. I'll try to get them out soon - not enough hours in the day.

In the meantime, I remain very busy with xVM Server work, and my sideline project of getting ZFS Automatic Snapshots into Solaris is hopefully just about done - the status on that, is that I putback the zfssnap role last week to the ON source tree, and the Desktop consolidation have delivered SUNWzfs-auto-snapshot to the WOS already, so we'll have the ZFS Automatic Snapshot service in nv_100 - w00t :-)

Back to the day-job: one of the things that popped out of the gate work for the xVM Server product, was a growing frustration with the older version of hg we're using to manage the xVM and xVM-Server gates. We need support for webrev -r to be able to produce readable webrevs for source trees being managed by Mercurial MQ, and had been maintaining our own private copy of hg and the cdm module for a long time.

We're still not quite at the level where we can move off it entirely, but I was able to spend some time alleviating the problem by a quick bit of ksh hackery. The attachment I sent to the scm-migration-dev mailing list, patch-webrev.ksh makes for slightly nicer webrevs of MQ patches. So, no longer do you have to try to get your head around diffs of diffs. (yuck!)

I suspect this script won't scale for very large source trees, but it's certainly a step in the right direction. Hope you find it useful.

Update (later that day): johnlev spotted a bug in my script where it was reporting more changes that had actually been made in the patch. I've got a new version here which does the trick. I've fixed the link above too.

I'm probably the last person on earth to discover this - but just today, I used the Mercurial bisect command, and thought I'd write up my experiences in case anyone else hasn't played with it before. I'd read about hg bisect in the hgbook, but never had an opportunity to use it in anger.

Here's the problem I was seeing - in builds of xVM Server that I've been doing, we were producing ISO images, but after installation, the pkg command wasn't working properly. Exploring the image a bit, with some help from the pkg python stack trace, I found the problem was that some items in /var/pkg were symlinks pointing to a non-existent mountpoint on the installed image.

Looking at the build logs from distro constructor, cpio was complaining that there was no space left on the device it was writing to. Digging around a bit more and running another build just to make sure, I found the source of the problem - we were df'ing the source directory for the cpio, then doing a mkfile of that size, creating a lofi device that big, then creating a UFS filesystem on that device. There was the problem - the space overhead incurred by the filesystem meant that we were trying to pour a gallon into a pint pot.

So - I knew what the problem was, pulling the tip changeset from the distro constructor even showed me that the problem was already fixed (my favourite kind of bug!) - the fix being to make the file which backs the lofi device just a bit bigger. My question was, what changeset introduced this fix? Enter hg bisect.

With it, you just need to identify where you know the code is bad, and where you know the code is good, and a test to determine whether the change is present. In my case, the test was really short:

grep "Add 1%" build_dist.lib

- but you could conceivably have the test build an entire OS image, install it, and check for the change. The bisect command then does a chop through all of the changesets, narrowing down to where the change was introduced.

In my case, a source tree of 105 changesets resulted in my only having to perform 6 tests to determine where the change occurred. A grep across 105 files would have completed in no time, but had I actually needed to build an OS image for each test, 105 builds would have taken a very long time indeed.

Here's some edited highlights:

timf@haiiro[435] hg bisect -g tip
timf@haiiro[436] hg bisect -b 0  
Testing changeset 52:42e67ad1e103 (105 changesets remaining, ~6 tests)
125 files updated, 0 files merged, 5 files removed, 0 files unresolved
timf@haiiro[438] grep "Add 1%" build_dist.lib   
timf@haiiro[439] hg bisect -b
Testing changeset 78:76e8ef490770 (53 changesets remaining, ~5 tests)
119 files updated, 0 files merged, 95 files removed, 0 files unresolved
.
.
timf@haiiro[447] hg bisect -b                
Testing changeset 103:b8d33c12a531 (4 changesets remaining, ~2 tests)
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
timf@haiiro[448] grep "Add 1%" build_dist.lib
	# Calculate the size of the pkg data directory.  Add 1% of the
timf@haiiro[449] hg bisect -g                
Testing changeset 102:ef08a25b1d1c (2 changesets remaining, ~1 tests)
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
timf@haiiro[450] grep "Add 1%" build_dist.lib
	# Calculate the size of the pkg data directory.  Add 1% of the
timf@haiiro[451] hg bisect -g                
The first good revision is:
changeset:   102:ef08a25b1d1c
user:        Karen Tung 
date:        Wed Aug 06 20:22:36 2008 -0700
summary:     2810 pkg archive size not big enough sometimes

So - I need to update our copy of distro_constructor to be based on changeset ef08a25b1d1c, which gets me the fix for 2810. Yahoo!

I stumbled on this tip today on planet.gnome.org about how to tune what gets displayed on your favourite planet - this has made me extremely happy, as I now get to have a userContent.css file that says:

@-moz-document domain(planet.opensolaris.org) {
  div.observatory div.person-info { display:none; }
  div.observatory div.post { display:none; }
}

Why? Well, in it's own words, "Planet OpenSolaris is a window into the world, work and lives of OpenSolaris hackers and contributors." - more particularly, I don't feel it's the right place for documentation or marketing spiel about OpenSolaris - there's other places for that.

Don't get me wrong, I'm thrilled that those guys are writing content about OpenSolaris that Google will cache and end-users will benefit from - they're doing a fantastic job! Personally though, I go to planet.opensolaris.org to read what people think: I don't go there to read software documentation or watch hundreds of screen shots of installation wizards, let alone read about quintuple-boot setups (gak!). Come on guys - there's got to be real people behind the marketeers?

So, to paraphrase Lt. Ripley, I vote we take off and userContent.css those entire posts from orbit. It's the only way to be sure.

Of course, I could be wrong - if so, feel free to load your text editor of choice, and with feeling, type div.timf div.person-info ... I'll totally understand!

There's two events coming up pretty soon that I wish I was able to attend:

Things have been pretty busy for me for the last few weeks. I've been helping with ZFS testing for putbacks to ON providing root filesystem support, and working on an ongoing putback for the same thing to the Install consolidation, helping test some backports of ZFS functionality to the next S10 update, and getting the opportunity to explore some new work stuff (which I'll talk about more in a future post)

While all this has been going on, we've been trying to get ourselves organised for an upcoming 2 week family vacation - which means I'm going to miss both OpenSolaris Developer Summit and CommunityOne: both opportunities to meet up with lots of the OpenSolaris folks - going to miss you guys! The summit last October was a blast, and this one sounds like it's going to be as good if not better, oh well.

On the plus side, I'm really looking forward to getting away for a while - it's my sister's wedding in Spain in May, and we're heading off on Saturday to stay over there for a while and generally get some much needed downtime. Haven't decided if I'm bringing a laptop or not, but leaning towards not at the moment!

Along with lots of QT with the missus and Bananas, I'm also looking forward to hanging out with the rest of the family, Gman included, who's jetting over from New Zealand, via the aforementioned conferences.

So, while I do wish I was there, I'm also glad I'm not (in a weird kind of way) - but am very happy that much of the summit will be recorded, so hopefully I'll be able to catch up when I'm back.

Just posted this to IE-OSUG mailing list:

We had a bit of a scheduling mishap with last night's meeting, which resulted in us deciding to cancel the meeting - we did hang around in the lobby of DIT till about 19:30 in case anyone else showed up, but my apologies if you missed us.

The good news, is that you still have a chance to catch Rafael's talk on NUMA machines and what was done in OpenSolaris to optimise performance on them. We're going to reschedule the meeting to a later date, but I'd like to try to find a time that suits as many people as possible.

On a more general note, a question to our members: how often would you like us to hold meetings: are we just killing ourselves trying to hold monthly meetings, or would less frequent meetings be better ?

Would you prefer more podcast content and fewer face-to-face meetings ? Are the topics that we've been doing of interest to our membership ? Are we publicising the meetings enough [ clearly not, given 5 people showed up last night, including the 2 speakers ]

Hopefully if we can get answers to the above, we might be able to inject more life into the user group! I'd really appreciate feedback that people might have.

cheers,

tim

Short notice here, but as DIT closes early during the Easter break, we're going to reschedule the next IE-OSUG meeting to Tuesday the 1st April so that we don't have to start/finish the meeting early.

I've updated the meeting page here and have updated the dates on the poster below. Looking forward to seeing you all on Tuesday.

There's an interesting discussion going on at the moment in the OpenSolaris Advocacy community, regarding a new logo for OpenSolaris.

Everyone has an opinion - the logos Glynn's been posting aren't bad (I like the "Circles" one, but not the "O s" ones) Lots of people have been posting their own logos for the project and have been commenting about some of the ideas out there - it's hard to say whether all this is a done deal, or whether there's still time for better logos to emerge.

Art is hard, and logos are even harder, I can't even begin to imagine the aesthetic, cultural and legal aspects of designing a good logo (not to mention the technical ones!) - and in the meantime, there's a maelstrom of ideas out there, happily churning around.

So, to further muddy the waters and mix metaphors, here's another to add to the pool, which I prefer more than any of the current logos (as if my opinion matters! :-) I initially wrote about this idea here.

Comments welcome here, but you're far better off joining advocacy-discuss and contributing there. I've posted a link on that mailing list to the svg source of the image above, so while opinions are good, actual art is better!

We're happy to announce the 18th Irish OpenSolaris User Group meeting! Rafael Vanoni is giving a talk about the changes that were made to OpenSolaris to support NUMA architectures. More information about the meeting, and a copy of the poster above is available on the Irish OpenSolaris User Group project pages.

If, after reading this post, you end up spending the rest of the day humming this then I'm not responsible, okay! You read the post, it's your fault! :-)

I'm looking forward to the upcoming OGB elections and will be interested to hear what the candidates up for election believe are the issues and concerns facing the project and community as a whole. It'll be good to listen to Barton's interviews of the candidates also.

I had a few moments the other day to come up with a new graphic for the OpenSolaris front page to further publicise the election - and, given that I'd done the previous "open(2)" logo, you can tell the amount of creativity that went into thinking up this one - perhaps I'll just stick to computers, I'm not sure graphics-design is my strong point! Thanks Derek for fixing the fonts, and adding the extra text.

(Trivia: the code we've used for the various bits of OpenSolaris art produced so far, comes from turnstile.c, zio.c, open.c, audiohc.c and poll.c - am I missing any?)

I've finally recovered enough to talk a bit about how we got on at FOSDEM last weekend.

All in all, it was a really hectic few days, getting up at 4am on Friday morning to get to the airport to make the red-eye flight to Brussels, and then coping with the hour time difference between Dublin and Brussels during a trip that had us burning the candle at both ends (getting up at 6:30am was really getting up at 5:30am on Ireland-time, still I've got a 17 month old daughter, so that prepared me for it a bit!)

The OpenSolaris stand was staffed by: (alphabetically)

• Casper Dik
• Sara Dornsife
• Tomas Dzik
• Patrick Finch
• me
• Giles Gravier
• Menno Laggeman
• Petr Sumbera
• Joep Vesseur

We also had special guest appearances from Steve Lau, Adrian De Groot, Jan Schmidt and Michal Bielicki - thanks for the help guys!

Our stand was a table, with a few chairs, a large OpenSolaris banner in the background, a few boxes of OpenSolaris starter kits, lots of copies of the Student guide (part of the "An Introduction to Operating Systems, A Hands-On Approach Using the OpenSolaris project" book) and not enough copies of the current OpenSolaris Developer Preview 2 LiveCD!

In terms of hardware, here's what we had with us, mostly brought from the Sun Netherlands office by Joep and Casper:

• 1 x Ultra 40 (with the metal side of the case removed, the inner plastic cover ensuring both proper airflow and a lovely view of the hardware) - this was our Sun Ray server for...
• 3 x Sun Ray 2FS units - 2 were at our booth, one beautifully painted with the KDE logo, and the matching KDE-skinned unit was over at the KDE booth (so we could do some demos of hot desking via a very very long ethernet cable, snaking around the corridors of the university where the conference was held) Joep and Casper did most of the work to get the Sun Ray server up and running, with Adrian providing the KDE goodness.
• 2 x 20" flat screens
• 1 x 15" MacBook Pro running SXCE on bare-metal and OpenSolaris Developer Preview 2 in a VirtualBox (or was it a xVM domU?)

• 1 x EEE PC also running OpenSolaris Developer Preview 2 (which I'm really glad I managed not to break during the course of the weekend!)
• Assorted other laptops floating about.

Between the Ultra 40 with it's innards exposed, the beautifully skinned Sun Ray clients and my tiny Eee PC (which now runs Compiz very smoothly indeed!), our stand had a lot of things to pull in quite a crowd - indeed we caused minor traffic jams in the hall from time to time.

Common questions asked by visitors to our stand were:

• "An Eee PC - wow, where did you get it!?" (followed by me giving a quick run-down of it's hardware, and then pure astonishment when they were told that we were actually running OpenSolaris on it, booting a ZFS root from a 4gb SDCard - cue ZFS demo)
• "What's inside those thin clients?" (followed by amazement when we did the hot-desk demo, including redirection of YouTube audio from one Sun Ray to another)
• "Is that OpenSolaris running on a Mac Book?"
• "Why should I run OpenSolaris instead of Linux ?"
• "Which version of OpenSolaris should I run ?"

All of which gave us ample opportunity to enthuse about our project, point out why we thought it was worth a try, find out what people liked and disliked about the project and admit to things we agreed needed work. Pointers to the OpenSolaris New User FAQ and the Unix Rosetta Stone were also well received.

ZFS and DTrace were obvious attractors for people coming to the stand - the down side was some occasional negative feedback about the opensolaris.org website, mainly confusion as to which distributions people should be running - we need to work on that imho.

In general though, I felt pretty good about being at FOSDEM - I learned a lot by having to talk about our OS on many different levels, everything from discussing the design of ZFS, to one nice person asking what sort of tools were available on OpenSolaris that a Windows administrator would be able to use (I gave a quick demo of Webmin - which looks really nice these days)

We even had some folks from a Belgian TV station who were brought to our stand by the FOSDEM organisers looking to do a short piece about my Eee PC - I showed them OpenSolaris running on it: they asked if I could boot the factory-installed Linux on it instead, which I declined politely, stating that OpenSolaris was a free and open source operating system too, and with a quick reminder that it also had an easy-to-use interface, they were happy to talk to Gilles (who speaks French) on TV - I just wish the "<i boot>" OpenSolaris stickers were a little larger!

Of course, it wasn't all work - after 6pm we were able to take in some of the nightlife of Brussels, and as a beer aficionado, I was in just the right place! We also got to meet up with Simon and some of the other Sun folks over for the conference: good beer and great conversation, always a nice combo.

Finally - I definitely need to praise the hard work of Patrick for organising our presence at FOSDEM, we really couldn't have done it without his help, and I was absolutely delighted to have been invited over to the conference.

Here's hoping OpenSolaris becomes even more popular by FOSDEM 2009, we've an interesting year ahead!

Just got back from FOSDEM this morning - far too tired to write up full account, except to say that it's an amazing conference, and I think the OpenSolaris stand was pretty popular. More to come as soon as I've had some sleep!

I'm heading over to FOSDEM at the end of this week and am really looking forward to it. My task? To be an OpenSolaris guy on the ground, help out at the OpenSolaris table, and answer any questions you might have about the project.

Of course, secondary objectives include mussels & chips with stevel and perhaps a bit of tourism around Brussels in the evening if there's any time to spare.

Hopefully, if I haven't broken it by then (not a bet I'd take, by the way) I'll also have with me my little Eee PC running OpenSolaris Developer Preview 2 - a 7" laptop, booting a ZFS root filesystem from a 4gb SD card.

If people had told me 10 years ago that we'd have a free (in all senses of the word) Solaris-derived OS running on one of these machines, I'd probably have asked who's round it was, but as I've said before, nothing beats a demo :-) If you're over at FOSDEM, drop by the OpenSolaris stand and say "hi!"

There's lots of traffic on ogb-discuss at the moment - both for and against the sentiment issued in Sun's responses to the OpenSolaris Trademark Questions.

As always, the excellent John Plocher has responded to several mails with a carefully considered post, which expresses how I feel about the whole thing particularly well - an excerpt:

As with any change, there are unanswered questions, uncertain futures, and activities that may become more difficult. There will also be opportunities and benefits that hopefully will outweigh those downsides.

If only we'd stop shooting ourselves in the feet.

John's original post is here.

I want to see more OpenSolaris distributions running on everybody's computers: helping to solve the trademark issue by collaborating on the Trademark usage and branding guidelines, along with resolving the parking lot items will let us progress towards that goal.

For the record, I'm still happy that Indiana is being called "OpenSolaris" not because it came from SMI, but because it looks like it's in the best position right now to get more people using computers running on the OpenSolaris codebase - if other distributions can match, or hopefully exceed that in the future, then the de-facto distribution will no longer be OpenSolaris, it'll be that distribution instead, whatever it's called.

Ultimately, I'm happy to let the users vote with their feet.

We held the 17th IE-OSUG meeting last Thursday - a reasonably successful night, despite everyone who got the train over from East Point arriving a bit late, myself included - sorry guys!

Mark Deegan, a lecturer from DIT kicked off the night with an ad-hoc talk about computing at DIT - he's trying hard to get more UNIX included in the computing courses there, and is doing a pretty good job seeding the department with people who have more UNIX skills. Getting more OpenSolaris in Universities is critical to the success of the project, so huge thanks to Mark for helping out here! Some of us pointed at the curriculum docs on opensolaris.org, so perhaps snippets from this course could be scattered throughout the various courses at DIT?

Mark also made a call for feedback for ideas on how they could further utilise their pretty extensive UNIX/Solaris/SunRay lab - maybe we'll see a few entries into the OpenSolaris Innovation Awards from DIT students as a result.

Next up, I gave a run down of the OpenSolaris news from December and January.

To finish the talks, Jan - who's recently joined Sun, gave an amazing demo of GStreamer managing a video feed from a webcam attached to an OpenSolaris laptop. The high point of the evening for me (and I think most people in the room!) was an impromptu performance of how GStreamer could be used to do live green screen effects on OpenSolaris. In my view, the best technology demos leave the audience grinning from ear to ear, and this was no exception - seeing Jan play a video in a colour-bound area, superimposed over a live feed from the webcam was really cool! I've since had a few people ask for the shell history of this talk (as you can imagine, it doesn't come over on the podcast very well!) - unfortunately, I wasn't able to save that, but Jan's written up his presentation in a blog post - which is well worth checking out.

We didn't have time to cover the last lightning talk - a review of ZFS since it integrated into the OpenSolaris codebase two and a half years ago. Perhaps if there's interest in that topic, we can cover it in a future meeting. Feedback welcome.

We finished off the night with some great food & beer over at The Bull & Castle, where the chat continued till closing time :-)

I've got the audio now added to our podcast feed - our typical bad quality audio is somewhat worse than usual this time around, I've done my best to reduce noise as much as possible, but apologies for any headaches this causes - we'll try to do better next time.

Last night/this morning Jim Hughes and others spoke at the Silicon Valley OpenSolaris Users Group about Indiana, the new packaging system (IPS) and gave some thoughts on the future of OpenSolaris.

The SVOSUG folks streamed the meeting live via UStream.tv, and if you happened to be awake, you could visit the streaming media site, and add comments in real-time - nice! As our user groups are so geographically dispersed, I'm really happy that Alan DuBoff and Bob Palowoda put in the work to better accommodate remote participants - US toll free numbers are only toll-free if you're calling from the States! [ Hey, perhaps the OGB could record their meetings this way ? ]

Anyway as it turned out, last night I had a minor case of insomnia and managed to catch some of the meeting, from about 4am GMT onwards. Thankfully though, anyone who got a decent night's sleep can still catch up, and view the recorded video of the session.

I'm more of an audio guy myself though, so I thought it might be worthwhile ripping the audio feed, and posting it for your listening pleasure. You can now find this in the Irish OpenSolaris User Group Podcast feed.

I haven't listened to the entire meeting yet, it's 2.5 hours long, but what I've heard is pretty interesting stuff. Hope you find this useful!

Pretty amazing stuff here from Adam Leventhal - a few months ago, I was really looking forward to seeing DTrace on OS X: it now sounds like the implementation of DTrace there has missed the point in some fundamental ways.

"Systemic tracing" to me means all of the system, not just the bits the lawyers are happy to let you look at. Here's hoping the Apple engineers get this sorted out.

[ Disclaimer: I haven't bought a copy of Leopard yet. I still find it a bit weird that I have to pay for something that includes ZFS and DTrace, when I can get working versions of them for free in my preferred operating system. If my Mac is still supported when 10.6 comes out, I might upgrade to that. For now, 10.4 does everything I need. ]

I just sent a mail announcement about our next IE-OSUG meeting, which is on Thursday January 31st 2008.

We haven't managed to scare up any main presenters this month, but that's okay - we've got a good deal of news to cover since we skipped a meeting in December, so we'll roll up news from Dec and Jan to begin with.

We're also going to try to run some Lightning Talks again - these worked pretty well the last time we did them, but they do really depend on volunteers to talk. It's pretty easy to talk for 5 minutes about something you care about - whether that's the latest feature in OpenSolaris, or some things that really bug you about the OS, or some ideas you have about user group activities, so hopefully we'll get a good range of speakers - it's up to you to participate :-)

Update 21st Jan: Please note the change of room for this meeting - it's now on in room KA:G-026, which is on the ground floor of the (newer red-brick) Annexe building at DIT Kevin Street. I've updated the IE-OSUG webpage as well as the poster that's attached to it.

Hurrah, my Eee pc arrived this afternoon - thanks gbax.com!. Getting my hands on one of these this side of Christmas was a challenge - Asus it seems just can't make them fast enough, but the excellent customer service from GBAX put other (larger) retailers to shame.

I haven't had much time to experiment around with the machine so far, but I have been able to plug an external disk with the Indiana preview on it, to see what's supported.

Out of the box, it boots as far as a gdm login screen, and with an external USB keyboard, I can get to a full desktop session. In general, it's pretty responsive and it didn't feel like I was running on a slow machine.

As suspected, the wired nic isn't detected, and neither is the wireless one - I'll post prtconf and scanpci output in a while, it might just be a case of tweaking driver aliases (since Solaris has support for the ath wifi card I think)

The laptop's keyboard also doesn't work on boot - from a bit of playing about, doing a modunload of kb8042 and modloading it again, followed by "devfsadm -i k8042" makes the keyboard work, but on rebooting, the same problem occurs. I haven't worked out what the problem is yet, but it's early days and I don't have time to experiment more tonight.

Otherwise, when running linux on the box, everything just works - the keyboard will take a bit of getting used to, but I wouldn't give up the form-factor of this little device!

Will post more if I manage to get more stuff working under Nevada.

Fantastic! Newboot Sparc putback last night - hearty congratulations to Jan and everyone else involved! Here's the flag day message.

I was on the testing team for this one, and it was a very long road from a testing perspective - we needed to verify that boot (and wanboot, jumpstart, cdrom and dvd booting) worked, that install and upgrade via GUI, ttinstall and jumpstart worked, that liveupgrade and flash archives worked and (more to the point) that every bit of sparc hardware we could lay our hands on came up without problems.

And, if that weren't enough, because there was code changes for x86 as well, we needed to make sure x86 booting and installation still worked too. The test matrix for this was absolutely massive as a result, and the testing was hard to automate.

We enlisted the help of testing groups right across the whole of the company to ensure we had the widest possible coverage of different machines, and this truly was a team effort.

If you're wondering why this is so important, PSARC 2006/525 explains more, but as well as ITUs (install-time updates) for sparc, this paves the way for ZFS Boot on sparc!

Last night, Michal gave a great talk & demo of IPS, the new package management system that's being developed for OpenSolaris.

In my book, nothing beats a demo, and Michal didn't disappoint - we got to see a package server being started, a package being created and published to the server, and finally being installed on an image, and refreshed along the way. I'm not entirely sure how well this will come over on the audio (haven't re-listened to it yet) but I'm sure Michal will be willing to answer any and all questions about it! Of course, if you're hungry for more pkg(5)-related stuff, you should also check out Stephen's talk, with accompanying audio - pkg(5) at the OpenSolaris Developer Summit.

I also gave the monthly run-down of OpenSolaris news - based on this list of links.

The audio from the meeting has been added to our podcast feed and we've also got a direct link to the mp3 for those so inclined.

I don't know if we'll hold a user-group meeting in December - the thinking being, that a lot of folks have pretty busy social calendars running up to Christmas, but please let us know via our mailing list if there's anything in particular you'd like to talk about or hear about. If there's enough interest, we might still hold a meeting, but do let us know.

I just posted an announcement for the next Irish OpenSolaris User Group meeting. This month, we're talking about IPS the new package management system being developed for OpenSolaris.

Michal Pryc has kindly volunteered to give us a quick introduction to it - so if SVR4 packages have been wrecking your head for a while, this might be just what you're after!

As usual, we'll also be doing the monthly roundup of OpenSolaris news, and of course will be adding the audio from the evening to our podcast feed. Feel free to download the poster above to help us publicise this event.

See you on Tuesday!

I'd love to try OpenSolaris on something like this:

More details here, here and here.

In terms of performance, while it doesn't appear to have the fastest CPU in the world, it'd probably be a lot faster than my 733mhz PIII, 256mb RAM GX110, which is able to run snv_70 with a pretty acceptable user experience for basic applications.

[ aside: I'm trying to get Indiana on the gx110 at the moment - I've booted the livecd to console mode, and am running the installer from a remote X session though. Getting the GNOME desktop to start took a loong time, and I ran out of patience waiting for the installer to load, but this machine is definitely not the target audience of Indiana, so I'm pleasantly surprised I've got this far - on more realistic developer machines, the installation experience is much much nicer. I'll add comments if the install finishes. ]

Back on topic - for the most part, the Eee PC appears to tick all the boxes for a laptop for me - small, light, reasonable battery life, but most especially, it's portable (though having not seen one in real-life, I wonder if the screen is too small?).

A 15" laptop like the Ferrari was a bit too heavy for the bike. I had been thinking about a Toshiba M500, having had happy experiences of an M100 a few years back (and am currently using a borrowed one with a cracked screen as a work-from-home desktop), but I've been holding off asking for a new laptop until something lighter than the Tecra M5 appears in our internal laptop purchasing system. If this Asus ran OpenSolaris though, it's nearly cheap enough for me to just buy one myself, rather than wait for the folks at CAMs to get them in...

Anyway, if anyone would like to send me a review model ( >= 512mb RAM if possible), I'd be more than happy to try OpenSolaris on it and report back my experiences! Based on the specs, does anyone know how much of the hardware is supported by OpenSolaris already ?

I last wrote about Solaris on the desktop nearly 3 years ago I think. Well, this evening, while browsing through blog referers, I came across this post, which I think is worth reading. Some positives, some negatives, but I particularly liked this part:

[ ... ] So what’s my point? The fact that I’m even bothering to voice these petty complaints about a Solaris system is fairly amazing. Think of all the things I’m not complaining about: X came up like magic, with the right screen resolution for my 24″ LCD panel. Audio works. Really - it just works. WIFI works, to my great amazement. I booted up, it asked me what wifi network to connect to, I provided it with my WEP key (which I have memorized, thankfully), and it just freaking worked. I’m still a bit stunned at that one. I mean, Ubuntu works that smoothly these days, but it’s still a pain in the ass to get Windows to work with the average wifi card if the driver for it isn’t included with Windows. Holycrap, Solaris beats windows at network configuration. Never thought I’d see the day. [ source: Wibblelog ]

Yes, there's still the naming thing, but all things considered, we're not doing too bad at all.