UbuntuIRC / 2011 /06 /08 /#ubuntu-cloud.txt
niansa
Initial commit
4aa5fce
raw
history blame
13.4 kB
=== dendrobates is now known as dendro-afk
[02:23] <maahes> is it possible to check out UEC on a thumb drive?
[03:03] <maahes> ping?
[03:03] * flaccid yawns
[03:08] <maahes> restate question since I got booted: is there like a live usb version of UEC?
[03:54] <flaccid> maahes: well you can run ubuntu from usb so no need for a diff version
[03:55] <maahes> flaccid: well what I mean, is I was looking at the documentation for UEC and it said 40gb minimum.
[03:55] <maahes> and I have her...a 4gb flash drive.
[03:55] <flaccid> that may not be enough, give it a go..
[04:07] <whytehorse> Is this the appropriate channel to find help with Ubuntu on Amazon cloud?
[04:09] <smw> whytehorse, yes.
[04:09] <smw> whytehorse, although, you were probably looking for a second opinion :-P
[04:10] <smw> whytehorse, the topic should give you a good idea of the topic of the channel.
[04:11] <whytehorse> great! I am having a lot of probs with amazon, specifically creating an ami... I tried the instructions from ubuntu but it fails with some error about the kernel is not available
[04:11] <smw> whytehorse, how are you trying to make the AMI? Are you basing it off a previous AMI?
[04:12] <whytehorse> I downloaded the image from ubuntu for uec 10.04
[04:12] <whytehorse> then I bundle it, upload, register, etc
[04:12] <flaccid> make sure you are specifying the correctiong --region
[04:12] <flaccid> us-east-1 is default
[04:13] <whytehorse> my bucket is on us-east-1
[04:14] <whytehorse> I mean I can complete the process and launch an instance but it crashes because it's loading the worng kernel... so I tried to specify the vmlinuz kernel and amazon doesn't like that
[04:15] <smw> whytehorse, do you have an error message?
[04:15] <smw> whytehorse, a link to the instructions you are following?
[04:16] <whytehorse> yes, libudev: udev_monitor_new_from_netlink: error getting socket: Invalid argument
[04:16] <whytehorse> mountall:mountall.c:3204: Assertion failed in main: udev_monitor = udev_monitor_new_from_netlink (udev, "udev")
[04:16] <whytehorse> init: mountall main process (693) killed by ABRT signal
[04:17] <flaccid> you either use a published AKI or a pvgrub AKI if your AMI is designed for grub boot
[04:18] <whytehorse> I am concerned that the published aki are old fedora kernels, is that true?
[04:18] <flaccid> the ones amazon publishes are
[04:18] <flaccid> fc8 usually
[04:19] <flaccid> you do know what ubuntu publishes official AMIs ?
[04:19] <whytehorse> yes, we tried some of those but they have issues with init consuming 300mb ram
[04:19] <flaccid> ubuntu/cloud has a bug tracker
[04:19] <whytehorse> the RightImage ones I think
[04:20] <flaccid> thats RightScale, not official ubuntu/canonical
[04:20] <smw> whytehorse, cloud.ubuntu.com/ami
[04:20] <smw> whytehorse, those are official.
[04:21] <flaccid> whytehorse: what was the AMI id you used for the RightImage? i can see if this is a currently supported image
[04:22] <whytehorse> RightImage_Ubuntu_10.04_i386_v5.5.9.1_EBS (ami-0c4ca665)
[04:23] <flaccid> whytehorse: that image is deprecated. for the latest RightImages, please see http://support.rightscale.com/18-Release_Notes/OS_and_Software_Package_Support#RightScale_MultiCloud_Images and http://oss.rightscale.me/
[04:27] <whytehorse> It WORKED!!! Sweet!
[04:27] <whytehorse> thank you!!!
[04:27] <whytehorse> haha I got kernel 2.6.32 :0
[04:29] <flaccid> using the latest/current images is a general rule of thumb in the cloud
[04:31] <whytehorse> am I stuck with this kernel forever or can I set up a server which updates the kernel through apt?
[04:31] <flaccid> if it is pv grub, you can install whatever linux-image and reboot
[04:32] <flaccid> if its not a pvgrub aki, you can't do any kernel changes
[04:33] <whytehorse> oic... which one is a pvgrub aki? It doesn't tell me on that site you gave
[04:33] <flaccid> there are many
[04:48] <whytehorse> is ubuntu just using the fedora kernel?
[04:49] <flaccid> no, the latest/natty should be using pvgrub and thus the native kernel
[04:49] <flaccid> you can check in console output
[04:49] <flaccid> or just look at the name/desc of the AKI..
[04:49] <flaccid> and of course uname -a
[04:50] <whytehorse> ok, so there's something taking up 400MB ram, I believe it's console-kit-daemon
[04:51] <whytehorse> er 117MB
[04:51] <whytehorse> and syslog is 127MB
[04:53] <whytehorse> on a micro instance that only leaves 200mb for other stuff to run on
[04:58] <flaccid> could just be memory caching
[04:58] <flaccid> but yeah ubuntu is generally bloaty like that
[05:06] <whytehorse> hmmm, I do realize memory is cheap but I can't imagine 400MB used up on a command-line server...
[05:08] <whytehorse> is this a bug in the amd64 server?
[05:09] <flaccid> probably; ubuntu is pretty buggy
[05:10] <flaccid> i'd be interested to see how you go on debian with one of these http://support.rightscale.com/21-Community/RightScale_OSS#Debian_GNU.2fLinux_MultiCloud_Images
[05:23] <whytehorse> weird, I get the same 400MB used on Debian too
[05:28] <flaccid> probably quite normal. linux kernel takes advantage of memory via caching to maxmise performance from available memory
[08:18] <whytehorse> looks like the 32 bit version uses only 180MB ram
[08:28] <flaccid> thats due to address space
[08:28] <flaccid> not sure why you are still measuring things in this way
[09:14] <kim0> Morning everyone
[09:20] <whytehorse> I'm just trying to figure out how much ram to allocate when creating servers
[09:21] <whytehorse> so it's important to know that 400MB is used by the system
[09:26] <koolhead11|afk> hi all
=== koolhead11|afk is now known as koolhead11
[10:07] <koolhead11> kim0, :D
[10:07] <kim0> koolhead11: hey man
[10:07] * koolhead11 is very happy
=== daker_ is now known as daker
[10:41] <mmarschall> Hi, I try to use the 32-bit ebs AMI in eu-west-1 region (ami-4090a634) with a custom kernel. Adding the generic-pae kernel to /boot/grub/menu.lst and rebooting hangs the machine. I found that the ami is using aki‐4deec439, which is said to be the right one for instance storage and aki‐4deec433 should be used for ebs. Might that be the issue? Thanks...
[10:49] <kim0> wonder why do you need the custom kernel
[10:50] <mmarschall> I want to use ksplice uptrack for rebootless kernel upgrades. They do not support ubuntu 10.04 LTS :-(
=== dendro-afk is now known as dendrobates
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== koolhead11 is now known as koolhead11|afk
=== dendrobates is now known as dendro-afk
=== daker is now known as daker_
[19:02] <kim0> = Start Ensemble cloud meeting -
[19:02] <kim0> Hi everyone
[19:02] <kim0> Got a few updates
[19:02] <kim0> from my side
[19:02] <kim0> Going quickly .. We now have a tutorial for authoring a formula Yaay "https://ensemble.ubuntu.com/docs/write-formula.html"
[19:03] <niemeyer> Woohay!
[19:03] <kim0> Going through this a few bugs were opened and squashed
[19:03] <kim0> So Ensemble now has much better debug-hooks functionality
[19:03] <niemeyer> kim0: Yeah, seriously.. thanks a lot for your help feeding the issues back onto the team
[19:03] <kim0> tmux is now being used instead of screen, things seem to work reliably
[19:04] <kim0> Also recovering from install hook errors are now handled in a better way
[19:04] <kim0> On a smaller scale, the mysql example formula was fixed temporarily at least, to offer login details of already created DBs
[19:04] <kim0> On another note
[19:05] <kim0> Clint published a very good introductory article on the topic of What is Ensemble
[19:05] <kim0> http://cloud.ubuntu.com/2011/06/so-what-is-ensemble-anyway/
[19:06] <kim0> I pushed it to multiple online venues .. it made its way to a few smaller online venues however unfortunately, it didn't get through most of the big ones .. I will retry with different tactics however
[19:06] <kim0> Other than that .. I've started reaching out to the MediaWiki community
[19:06] <kim0> since that's the one where we have fairly good formulas
[19:06] <kim0> posted today on MediaWiki forums and mailing list
[19:07] <niemeyer> Oh, nice
[19:07] <kim0> demo'ing quickly what Ensemble is capable of
[19:07] <niemeyer> kim0: What was the reaction so far?
[19:07] <kim0> and linking to Clint's article and the main project page
[19:07] <kim0> well it was a couple of hours back .. only a single reply so far adding a few links
[19:07] <kim0> I will keep beating the horse however (in a good way)
[19:07] <kim0> :)
[19:08] <niemeyer> kim0: Cool :)
[19:08] <kim0> um .. that's mostly it for me on my side .. Any cool code updates? I see some bugs about resolving deps and ec2 firewall ?
[19:09] <niemeyer> kim0: Hmmm
[19:09] <hazmat> kim0, the open/close port work and service-config are ongoing but nearing completion
[19:10] <kim0> Very cool
[19:10] <hazmat> kim0, the autoresolving dependency prototype was merged for a historical artifact, but removed pending work on the repository (the prototype was against a local repository directory)
[19:10] <niemeyer> kim0: Yeah, I think next week will be more exciting from a news perspective
[19:10] <kim0> a ha
[19:10] <niemeyer> kim0: Unfortunately the few things we have in progress still haven't landed
[19:10] <kim0> Ok great
[19:11] <kim0> One final point I'd like to discuss
[19:11] <niemeyer> Ok
[19:11] <kim0> is having a running list of bitesized bugs ..
[19:11] <kim0> I'll repaste from #ubuntu-ensemble
[19:11] <kim0> I'll start a campaign to go out and get people interested in contributing to Ensemble. For that, we'd probably need a list of bitesized bugs that we can point people to
[19:11] <kim0> This has worked well for Unity, so hopefully for us as well
[19:11] <kim0> so we'd need a tag a few bugs as "bitesized". Or maybe file new bugs if appropriate. This can be easy to fix bugs, or nice to haves that are not too uegent for the team and are low barrier for a newcomer
[19:11] <kim0> if we can have 10-15 bugs tagged/created, that'd be awesome
[19:11] <kim0> I'll probably file some new bugs for important formulas I'd like to see written, and I'll pimp them to the respective communities. But where I need the help most, is code bugs (improve test coverage, fix a small bug ..etc)
[19:11] <kim0> that's it .. what do you guys think
[19:12] <hazmat> kim0, so effectively a level of effort/difficult tag on the bugs for bug 'bites'?
[19:12] <hazmat> s/difficulty
[19:12] <kim0> hazmat: yes effectively something you can resolve in less than an hour :)
[19:12] <kim0> it's a learning on-ramp for newcomers
[19:12] <niemeyer> kim0: That sounds reasonable
[19:13] <kim0> improving the test suite .. better docs .. added minor functionality ...etc
[19:13] <kim0> all good targets
[19:13] <niemeyer> kim0: Yeah, we have quite a few things which are not really big issues at the moment
[19:13] <niemeyer> kim0: If you find someone who'd be interested, please point us towards the IRC/ML
[19:13] <niemeyer> kim0: We can easily pin out tasks
[19:14] <kim0> niemeyer: hmm yeah sure will .. but it'd be great to have the written list of bugs
[19:14] <kim0> since just viewing them, can make someone say "I can do that"
[19:14] <niemeyer> kim0: We have a list of bugs:
[19:14] <niemeyer> http://ensemble.ubuntu.com/kanban/dublin.html
[19:14] <kim0> yes, we just need to tag the easy ones with bitesized
[19:14] <kim0> :)
[19:15] <kim0> so we always have a running list of newcomer friendly bugs
[19:15] <niemeyer> kim0: No need for tagging at this point..
[19:15] <kim0> what do you think so
[19:15] <niemeyer> kim0: E.g.
[19:15] <niemeyer> "No README"
[19:15] <niemeyer> kim0: It's there, in the list
[19:16] <kim0> Ok better deal
[19:16] <niemeyer> kim0: Test should put zk log to /dev/null or tempfile by default
[19:16] <niemeyer> ensemble terminate-machines should have an --all-unused option
[19:16] <kim0> I'll try to tag what I think looks easy to me .. then someone reviews ?
[19:16] <niemeyer> kim0: Etc..
[19:16] <niemeyer> kim0: If you find people, we have the bugs ;-)
[19:16] <kim0> I want to have a list to show off .. that itself helps
[19:16] <niemeyer> kim0: What I mean is that I don't think a process for this is needed, yet
[19:16] <niemeyer> kim0: Find the people, and we create a process around it
[19:17] <niemeyer> kim0: I've just pointed out three bugs above..
[19:17] <kim0> Yeah, I'm sure you know all of em
[19:17] <niemeyer> kim0: We all do.. the URL above has the full list
[19:17] <kim0> the easy ones I meant
[19:18] <niemeyer> kim0: Ok, hold on
[19:18] * kim0 holds on
[19:19] <niemeyer> kim0: #684573, #781920, #754318, #716401, #769036, #764938
[19:19] <kim0> oh that's fast!
[19:19] * kim0 hugs niemeyer :)
[19:19] <niemeyer> kim0: Let's get these fixed.. and then we look at a process
[19:19] <kim0> Awesome that's all I asked for
[19:19] <kim0> Great stuff .. that should be about everything
[19:20] <niemeyer> kim0: Cool, that's a deal then :-)
[19:20] <kim0> Thanks everyone
[19:20] <kim0> o/
[19:20] <kim0> enjoy hacking
[19:20] <niemeyer> kim0: Thank you!
[19:20] <kim0> = Meeting end =
[19:21] <kim0> If anyone is around looking at this, not sure what this is about .. ping me immediately :)
=== dendro-afk is now known as dendrobates
=== dendrobates is now known as dendro-afk
[23:00] <d122> was there an ensemble class?
=== dendro-afk is now known as dendrobates