[00:36] kirkland, im thinking on removing the release codename from the config and to only obtained from the cache file where Im storing it. What do you think about this? [00:39] and to load all the defaults from config file, instead of setting defaults where not set in the code === dendro-afk is now known as dendrobates [02:08] funkyHat: you still around? [02:08] MTecknology: just a bout [02:08] s/a b/ab/ [02:09] funkyHat: I'm getting this now - http://dpaste.com/180694/ [02:09] MTecknology: ok [02:09] That looks fine to me [02:10] funkyHat: oh.. [02:10] OH! [02:10] MTecknology: is it working? ⡈) [02:11] I sent the message from the wrong addy - that's why the message isn't going through [02:12] hehe [02:13] funkyHat: DONE! [02:15] funkyHat: How can I change the mailman admin email addy? [02:16] MTecknology: can't remember, one of the helper scripts can do it [02:17] MTecknology: oh, I think it's an option in the web admin interface [02:17] in general options [02:17] funkyHat: where's the actual admin interface? [02:18] http://lists.kalliki.com/mailman/admin/test [02:23] funkyHat: oh, I thought there was one interface for more admin [02:27] funkyHat: so how can I delete/add lists? [02:28] funkyHat: I need to recreate the mailman mailing list - i muffed it up some and that's probably easiest - no need for test either [02:32] kirkland, were you able to seen the messages i left you? [02:35] RoAkSoAx: yeah ... so I was thinking about this more [02:36] RoAkSoAx: we should talk about your graphical front end [02:36] RoAkSoAx: because i think that's going to make it a *lot* easier [02:36] RoAkSoAx: basically, I suggest that we change s/lucid/maverick/ in the Lucid code just before Lucid releases [02:37] RoAkSoAx: i think the gtk frontend should make the selection of the ubuntu codename *really* easy [02:38] kirkland: so powernap [02:38] kirkland, i was planing to merge my code to obtain the codename into the PyGTK but, still, i would have to hardcode the CACHE if we keep how the configgile is loaded now [02:38] lifeless: yessir [02:38] kirkland: is there any reason what it doesn't isn't on by default all the time ? [02:38] does it have downsides on laptops? [02:38] lifeless: i run it on all 5 of my mythfrontends [02:39] lifeless: arguably, it duplicates some of the behavior you get (or can control) with gnome-screensaver and the gnome power utilities [02:40] lifeless: obviously we don't have those in the Ubuntu Server, so we wrote our own lightweight screensaver/powersaver daemon [02:40] kirkland: hmm [02:40] I think where i am coming from is this: [02:40] lifeless: as it's written right now, powernap simply watches the process table for the presence or absence of a specified list of processes you want to monitor [02:41] - only using power when you have something to do [02:41] - we do things we don't want to do [02:41] (such as npviewer.bin never freaking stopping) [02:41] lifeless: ie, the processes that are running if your system is "doing work", and if they're not running (and haven't been running for a configurable while), then your server is deemed inactive and can powernap [02:41] and pulseaudio not closign the sound card when there isn't sound to play [02:41] lifeless: heh [02:42] lifeless: right, right ... [02:42] * kirkland thought lifeless was going to come at it from a tcp port usage perspective :-) [02:42] (because that's something else that's been on my mind) [02:42] kirkland: sadly, my keyboard and mouse don't connect over the network [02:42] be nice to get them all PAN'd up. [02:43] lifeless: you should enable ssh :-) [02:43] oh, it is, it is. [02:44] kirkland, Anyways, I would love to resolve this so that I don' [02:44] don't have to worry about it later [02:44] lifeless: okay, go on .... [02:44] kirkland, would you prefer me to do it both ways so you can take a look at it and give your thoughts? [02:44] no thats about it [02:45] lifeless: stand by for one moment [02:45] just mulling on how to get less idiotic power wastage in our stack [02:45] RoAkSoAx: okay, let's not dwell on calculating the latest release any longer ... [02:45] like, epoll timeouts are daft. [02:45] RoAkSoAx: establish your cache [02:45] RoAkSoAx: and just make sure that the user can update that cache by selecting the release they want to testdrive through your GTK gui [02:46] RoAkSoAx: we'll handle the distro-default in the config file just as a parameter [02:46] RoAkSoAx: we'll SRU twice a year when sabdfl picks his funny adjective/animal combo [02:46] RoAkSoAx: but each user will pointy/clicky the release they want to test [02:46] RoAkSoAx: so the flow should look something like: [02:47] a) pick distro (Ubuntu, Debian, etc...) -- you work on Ubuntu, but make a framework so that others can submit patches to let this work with Debian, Fedora, and friends [02:47] b) pick a version (here's your codename list -- lucid, maverick, etc) [02:48] c) pick a flavor (desktop, server, netbook, mythbuntu, etc) [02:48] d) pick an arch (amd64, i386) [02:49] RoAkSoAx: kind of wizard like [02:49] kirkland, for each time the user logs in? [02:50] RoAkSoAx: and also an Edit -> Preferences menu with the rest of the configurable things in /etc/testdriverc [02:50] I mean launches the app? [02:50] RoAkSoAx: only if they choose "wizard" [02:50] RoAkSoAx: in the Edit -> Preferences tab the user can set their default $distro, $version, $flavor, $arch [02:51] kirkland, the preferences menu is already in my mind. The wizard cool idea. though I was kinda thinking the interfae differently [02:51] RoAkSoAx: okay ... what's your idea? [02:51] kirkland, will draw it and send it to you :) [02:52] RoAkSoAx: sounds good [02:52] RoAkSoAx: let's track it in a wiki or the blueprint [02:52] RoAkSoAx: keep as much of this public as possible [02:52] RoAkSoAx: i suggest a TestdriveGTK wiki page [02:53] kirkland, but is basically having different tabs like "Tab Ubuntu" "Tab Kubuntu" "Tab Edubuntu" "Tab Mythbuntu", and each tab separate by: "amd64" "i386" [02:53] RoAkSoAx: i'd also like to look into getting Testdrive offered as a handler for ISO files, if they're liveCD ISO files [02:54] RoAkSoAx: such that if you right click on an ISO in nautilus, one of the options is "Run this ISO in Testdrive" [02:54] kirkland, yeah Im planing to do that as soon as It's officially announced from the GSoC and I wanna come up with the drawings of the possible interfaces before making it public [02:54] RoAkSoAx: alrighty ... think about both [02:55] RoAkSoAx: i think the wizard approach should be relatively straight forward [02:56] kirkland, yeah wizard should be easier, but personally I would probably go directly to the main interface because I personally would hate to be using the wizard all the time [02:56] would be simpler to have the interface right there after launching the APP [02:56] RoAkSoAx: alrighty [02:57] RoAkSoAx: for the images you already have cached, yes, i agree [02:57] RoAkSoAx: perhaps the wizard is how you add a new iso to your cache the first time [02:57] RoAkSoAx: once you've run it once, it's cached, then it shows up on the default canvas as a button [02:57] RoAkSoAx: like a "recently used" document [02:58] kirkland, or maybe the wizard can be used whenever we select the tab "Others" [02:58] RoAkSoAx: okay, how about this ... let's run with the "automotive" theme of Testdrive [02:58] RoAkSoAx: your default canvas is your "garage" [02:59] RoAkSoAx: which shows the "vehicles" you have in your garage [02:59] RoAkSoAx: it's empty, initially [02:59] RoAkSoAx: when you want to "add a vehicle to your garage for testdriving", you go through the wizard [02:59] kirkland, such as add: Ubuntu, other "vehicle" Kubuntu [02:59] RoAkSoAx: once you've gone through the wizard once, and rsync'd an iso, it shows up as a "vehicle" in your "garage" [02:59] and so on [03:00] RoAkSoAx: a Vehicle is Ubuntu-Lucid-Desktop-amd64 [03:00] RoAkSoAx: or Ubuntu-Hardy-Server-i386 [03:00] RoAkSoAx: or Debian-Sarge-i386 [03:00] RoAkSoAx: or Kubuntu-amd64 [03:00] RoAkSoAx: whatever it takes to describe *one unique ISO* [03:02] kirkland, ok, I'll have to think that thru, however, my plan is this: Step 1: getting testdrive modularized (including those changes I was thinking of). Step 2: Version 1.x which includes the "simple" testdrive version. Step 3: Version series 2.x: Add tabs for other distros/Xubuntu versions (kubuntu, mythuntu), and Step4: Version series 3.x could be adding support to add different distributions, and etc etc [03:03] and adding wizard and stuff like that [03:03] cause I don't want to jump into everything from scratch because it would be painful [03:03] if we handle that with versions as Im proposing I do think would be better idea [03:03] RoAkSoAx: okay [03:04] is like "Adding Features" [03:04] per release [03:04] RoAkSoAx: okay, we'll need to set some primary/secondary/tertiary goals for the overall project, but i think something like this should work [03:05] RoAkSoAx: i hope that the modularization is about 10-15% of the overall work [03:05] kirkland, yeah I mean, we don't have to rush features right now. [03:06] kirkland, in my opinion modularization is around 60% [03:06] RoAkSoAx: hmm, then we need to discuss what you mean by modularization [03:06] RoAkSoAx: i can split it into a couple of functions in a matter of a few hours [03:06] RoAkSoAx: and then we can make it more object oriented [03:07] kirkland, im doing it object oriented right now [03:07] kirkland, lp:~andreserl/testdrive/module [03:07] RoAkSoAx: but i don't see that as taking 2 of the 3 months of this project [03:07] kirkland, files "testdrive", "testdrive.py" [03:07] kirkland, I was thinking modularization to be done by the end of may, mid june [03:08] RoAkSoAx: how long are you working on the project, overall? [03:08] RoAkSoAx: first glance, testdrive.py looks about like I expect it would [03:09] kirkland, well this week haven't work more than couple hours due to schoolwork, but i've worked like 10 hours a week for 2/3 weeks [03:09] kirkland, im planing to work on it 20+ hours a week when the program starts [03:09] RoAkSoAx: cool [03:10] since right now i'm fed up with schoolwork due to i;m in the last weeks of class, so I have papers, presentations, proejcts due [03:10] finals [03:10] too [03:10] RoAkSoAx: okay, so as by the time Maverick opens, I'd like to merge your modularized testdrive pretty much en masse [03:10] RoAkSoAx: sure thing, i understand ;-) wasn't too, too long ago I was in college :-P [03:11] RoAkSoAx: basically by UDS [03:11] RoAkSoAx: speaking of, are you coming to UDS? [03:11] kirkland, hahah yeah I'm sick of college, and Yes I would love to have the modularization done by the UDS so that we can start working in parallel without affecting each others work. [03:12] RoAkSoAx: great, yeah, so let's set a tentative goal to get the OO code merged by UDS [03:12] kirkland, and as of today don't know yet since I've to apply for Visa, will send my docs by friday [03:12] RoAkSoAx: gotcha [03:12] RoAkSoAx: we'll plan on holding a testdrive-gtk session at UDS (which you'll just attend remotely, in the worst case) [03:13] kirkland, yes I was actually thinking about that today, so I would at least want to have a simple interface for testdrive-gtk by the UDS to show [03:13] on that just create the buttons and be able to launch [03:13] RoAkSoAx: hopefully you and I will have hashed out the basic design before UDS, but we'll put that out for discussion at UDS [03:14] RoAkSoAx: oh, that's certainly not necessary [03:14] RoAkSoAx: screen shots or napkin sketches would be sufficient for UDS [03:14] RoAkSoAx: focus on the modularization/OO code before UDS; draw up your ideas on a napkin and scan it ;-) [03:14] kirkland, i already have an interface that creates the buttons, but haven't added the launch functionalioty yet... :) [03:14] RoAkSoAx: alrighty [03:14] RoAkSoAx: i think we're on the same page [03:15] RoAkSoAx: did you already create the blueprint? [03:15] kirkland, yeah, it is really simple though [03:15] RoAkSoAx: url? [03:15] RoAkSoAx: i'd like to capture a few of these points we just discussed in the whiteboard [03:16] kirkland, https://blueprints.launchpad.net/testdrive/+spec/gsoc-testdrive-modularization [03:16] kirkland, https://blueprints.launchpad.net/testdrive/+spec/gsoc-testdrive-frontend [03:16] RoAkSoAx: could you rename those to server-maverick-testdrive-modularization and server-maverick-testdrive-frontend ? [03:17] RoAkSoAx: actually, just the last one, server-maverick-testdrive-frontend [03:17] kirkland, what would help me know is if you could take a look to my modularization code so far, and give me your thoughts about it, to see if I'm going in the right path since, to be honest, is been a while since my last OO programming project [03:17] RoAkSoAx: we shouldn't need a UDS session on the modularization [03:17] kirkland, server-maverick or desktop-maverick? [03:19] RoAkSoAx: make it server-maverick for now ... it might get switched to the desktop team, but i'm on the server team [03:20] RoAkSoAx: and i'm reporting to my manager our GSoC effort [03:21] RoAkSoAx: as for your OO, like i said, first glance looks okay; i haven't run it yet [03:21] RoAkSoAx: i'm not a python OO expert either, so it might help to put it out for some peer review at some point [03:21] RoAkSoAx: heck, lifeless might even give it a look over at some point :-) [03:25] kirkland, haha ok cool. Anyways, just send you an image to your @ubuntu email with my idea of interface [03:27] RoAkSoAx: thanks [03:27] lifeless: alrighty ... so where were we .... yeah, so powernap .... [03:28] lifeless: i'm hoping for a UDS session with AmitK [03:29] lifeless: i'd like to teach powernap how to "do" and "undo" a configurable set of power optimizations when something "happens" or "doesn't happen" as the case may be [03:29] MTecknology: rmlist [03:29] kirkland, btw... registered the blueprint so i guess i'll have to wait for it to be approved. I'll take notes of todays conversation and put them in the whiteboard of the blueprints I already created [03:29] funkyHat: thanks for all of your help :D - This thing is pretty awesome :) [03:29] RoAkSoAx: i just updated https://blueprints.edge.launchpad.net/testdrive/+spec/gsoc-testdrive-modularization [03:29] MTecknology: and newlist or http://lists.kalliki.com/mailman/create [03:29] RoAkSoAx: added a couple of workitems, and took a note [03:30] MTecknology: no problem ⡈) [03:30] kirkland, awesome. will do [03:30] RoAkSoAx: https://blueprints.edge.launchpad.net/testdrive/+spec/server-maverick-testdrive-frontend-gsoc [03:30] RoAkSoAx: i renamed it for you [03:31] kirkland, shouldn't that be registered in Ubuntu for the UDS? [03:34] RoAkSoAx: i just proposed for the sprint [03:34] RoAkSoAx: we should try to get one or two of the Design/User-Experience team in that session! [03:35] kirkland, Ok will do. I';ll try to draw a better sketch and try to ping someone there eto see what they think [03:41] New bug: #557773 in php5 (main) "php5-cli fails apt install due to mislabeled libkrb53 dependency" [Undecided,New] https://launchpad.net/bugs/557773 [03:45] RoAkSoAx: thanks [03:52] kirkland, btw.. if you plan to run the modularized code you need to also copy testdriverc into /etc since I've changed it to be compatible with config parser [04:07] RoAkSoAx: still around? [04:07] kirkland, yep [04:08] http://people.canonical.com/~kirkland/testdrive-gtk.html [04:08] http://people.canonical.com/~kirkland/testdrive-wizard.html [04:08] RoAkSoAx: ^ [04:08] kirkland, awesome!! [04:13] RoAkSoAx: http://people.canonical.com/~kirkland/testdrive-gtk-2.html [04:14] RoAkSoAx: one more ^ [04:14] RoAkSoAx: that table should be sortable by clicking on any of the tabs across the top [04:14] RoAkSoAx: might have more fields, like size [04:14] RoAkSoAx: whatever [04:14] RoAkSoAx: anyway, that's my 10-minute sketch ;-) [04:16] kirkland, that's a good idea. I'll do improvements such as classify them into tabs (as my sketch). However, Im not sure if all that is achievable in 3 months. So, I would guess that we'll first have to define which features we would like to see and then, start implementing from that point [04:21] RoAkSoAx: agreed; let's use UDS as the pivot point by which we define and prioritize features [04:21] RoAkSoAx: up until UDS, focus on the foundation [04:22] kirkland, I'm, though I'm implementing simple pyGTK stuff to give me some ideas of how the modularization should go to be able to use it in both sides (command line/PyGTK) [04:23] RoAkSoAx: i see you dropped some of my comments from the whiteboard ... would you please establish a wiki page and move them there, if you don't want them in the whiteboard? [04:24] kirkland, did I? probably we edited the whiteboard pretty much at the same time and they got lost since I didn't delete anything :S [04:25] RoAkSoAx: http://pastebin.ubuntu.com/410843/ [04:26] kirkland, yep I did those changes when your comments weren't there so I guess we just found a bug in LP :) [04:27] RoAkSoAx: fair enough ... would you mind grabbing them from that pastebin and chunking them back in, pretty please ;-) [04:28] kirkland, done already :) [04:28] RoAkSoAx: rock on [04:28] RoAkSoAx: okay, i'm calling it a night [04:28] RoAkSoAx: later dude [04:28] kirkland, have a good one === erichammond1 is now known as erichammond === robbiew is now known as robbiew_ === erichammond1 is now known as erichammond [06:17] This karmic-server did not come up after reboot. After going through a cumbersome process of bringing a monitor in its range, I see that grub-2 menu has no timer, and hence one must hit the enter key to proceed. [06:17] Server was rebooted after 120 days and there were apparently 4 kernel updates in the mean time. [06:18] Is there a way to ensure that such problems don't occur on servers? [06:30] Searching the forums, it appears to be a common issue. [07:11] aaditya, grub by default not having a timer? That's odd [07:50] lukehasnoname: Based on certain forum posts, I realized that it was due to some failure during previous bootup. Still no clue what that problem was. [07:52] syslog did not help much and at this time I worked around somehow [07:52] I also verified that /etc/default/grub was in perfect shape [07:54] This is precisely the issue: http://ubuntuforums.org/showthread.php?t=1283800 [07:55] grub getting stuck on a server based on one issue is not reasonable in my understanding. It's potentially a huge headache for the sysadmins. [07:55] <_ruben> hmm .. i have a (long-running) script with output redirection to a logfile which filled up the disk, deleting the logfile wont free the space since the file is still in use, any ideas? [07:58] how do I install a package from lucid, on karmic? [07:58] specifically, python-pyinotify [07:59] <_ruben> the complex method would adding lucid repos + apt pinning, or just download the debs+deps from your favorite repo [08:01] thanks [08:36] New bug: #557890 in samba (main) "transfer lockup connecting to a NetApp/CIFS share" [Undecided,New] https://launchpad.net/bugs/557890 [08:43] aaditya: look here: bug 420077 [08:43] Launchpad bug 420077 in grub2 "grub2 has no timer" [Undecided,Confirmed] https://launchpad.net/bugs/420077 [09:36] New bug: #557924 in php5 (main) "php5-cgi crashed with SIGSEGV in zend_get_constant_ex()" [Undecided,New] https://launchpad.net/bugs/557924 [10:02] Hi there. I have a question regarding Ubuntu Enterprise Cloud. Which networking mode I need to configure when I have a dedicated subnet of 4 ips for each node in random ip ranges? [10:07] Hi, [10:08] I'm trying to upgrade an Ubuntu 8.04 server to latest 10.04 beta, but do-release-update -d keeps suggesting intrepid for upgrade [10:11] <_ruben> merlijn-: is your 8.04 server fully up to date? [10:12] <_ruben> merlijn-: and it seems you have it configured to not want to upgrade to lts versions only [10:12] yea I think I fixed it now [10:12] silly typo in /etc/update-manager/release-upgrades [10:39] alvin: looks like that's the issue I was dealing with. I've added a comment to the bug. [10:39] Thanks! === swift_ is now known as swift === swift_ is now known as swift [11:26] New bug: #558335 in clamav (main) "freshclam log reports warning about libclamav version" [Undecided,New] https://launchpad.net/bugs/558335 [12:22] hey folks, is there a resource which lists the latest kernel and ramdisk ids (aki/ari) for the official ubuntu/canonical ec2 images? preferably with their associated versions. [12:23] I compiled a list of what looks like the latest versions for the current LTS (Hardy) http://alonswartz.pastebin.com/H9CSdyUv but it would be useful to know their versions [12:36] updated the list according to include "official" by ami id's listed on the ec2 starter page [12:37] are these kernels/initrds built against the latest xen modules? [12:50] I've got a techie installing ubuntu ina remote location. He is failing at the raid setup screen. Can I ssh in and continue it for him? [12:57] the official ec2 hardy amis are 2010-01-28, but the latest xen modules 2.6.24-27 were released on 2010-02-04, does that mean the aki/ari's were compiled against 2.6.24-26 (released on 2009-12-04)? [13:06] smoser: howdy [13:08] Hi there. I have a question regarding Ubuntu Enterprise Cloud. Which networking mode I need to configure when I have a dedicated subnet of 4 ips for each node in random ip ranges? [13:11] kirkland: how did your all-night-cloud-image-testing go ? [13:11] ttx: not so well [13:12] I wish I could start instances in my cloud :p [13:12] ttx: only did 9 rounds [13:12] ttx: i think i overloaded it [13:12] it says I don't have ips [13:12] hmm [13:13] kirkland: right... a better test would be to run the with-ramdisk images in parallel with the no-ramdisk [13:13] ttx: i just changed that and started over [13:13] i don('t thin the with-ramdisk would better succeed in such a test tbh [13:13] binBASH: probably an issue with your range of public IPs [13:14] binBASH: for the first install of UEC I'd suggest following http://testcases.qa.ubuntu.com/Install/ServerUECTopology1 [13:14] ttx: Yeah, could be. I wonder how to configure those. My Provider gives me 4 ips / node. All not on the same ip range. [13:15] ttx: ill take the memcached one [13:15] if I run euca-describe-availability-zones verbose all looks fine though [13:15] binBASH: you can list them rather than using a range [13:15] A,B,C instead of A-D [13:16] zul: ok [13:16] ttx: Yeah, but how it will detect on which node the vm needs to be started? [13:16] zul: please make sure you can reproduce the issue first, so that you can confirm it's fixed by the patch :) [13:16] zul: also could you cover http://iso.qa.ubuntu.com/qatracker/test/3891 ? [13:16] update: nevermind, i found what i was looking for in the ubuntu-on-ec2 ppa [13:17] ttx: You know the ips are bound to node, and cannot be moved to another one. [13:17] binBASH: hm? The nodes from a given cluster should live on the same network [13:17] ttx: sure [13:18] ttx: So as workaround I would probably need many clusters? ;) [13:18] binBASH: you should run on a private network and properly NAT [13:19] so that you don't have this "4 IP per node" linmit [13:19] ttx: NAT would be not fine in this case, there is a traffic limit of 2 tb / node. [13:23] ttx: What I need would be, instances with ip 1.1.1.1-1.1.1.3 can be run on node A, 2.2.2.2-2.2.2.3 can run on node B etc. [13:23] :) [13:23] dunno if this is possible however [13:23] kirkland, here. whats up? [13:24] smoser: wanted you to check the log i sent you [13:24] smoser: see if you see any aberations [13:24] glitches in the matrix ;-) [13:25] kirkland, from last night ? [13:26] the one from last night looks reasonable. [13:26] smoser: yeah [13:26] smoser: my overnight run aborted [13:26] smoser: i tried to run 8 instances every time [13:26] smoser: and i don't think my cloud could handle that [13:26] smoser: within your logic anyway [13:27] ttx, i didn't bother trying beta 1 without ramdisk as i know it would fail. i didn't bother trying with ramdisk for current because i have no failures related to content of the image (only to the hypervisor) [13:28] kirkland, can you put that log out somewhere ? [13:28] smoser: we believe that those errors that Dustin and I hit are not linked to the with/noramdisk, but rather to the bad performance of our laptop-based cloud... but having some hard evidence that noramdisk doesn't perform worse than withramdisk would help :) [13:28] i'm just interested in seeing it. [13:28] it seems to me that eucalyptus will return success from euca-run-instances before it really has space, and when it does, gets into bad limbo [13:29] ttx, i believe we've seen zero failures that we can at all relate to the ramdisk-less image [13:30] so, there is no way that with-ramdisk image could perform better [13:30] smoser: ok [13:30] the only thing i've seen as a failure in the image is the metadata service was busted once for me [13:30] smoser: i've overwritten that log, but it's in my byobu scrollback [13:30] smoser: i'll copy and paste [13:30] kirkland, no worries. [13:31] i do think that the test case is legit though, and shows general failure of euca under load [13:33] smoser: http://pastebin.ubuntu.com/411023/ [13:33] smoser: but i agree with ttx, i don't think this is ramdisk related [13:33] smoser: well, general failure of eucalyptus on underpowered hardware, at least [13:34] kirkland: do you like my blog meme idea ? [13:34] ttx: sure, go for it [13:35] jiboumans: roaksoax has put together a couple of blueprints for the GSoC testdrive project [13:35] jiboumans: should we label those server-maverick-testdrive* and set you to the approver? [13:35] kirkland: sounds like a great idea [13:36] feel free to add -gsoc- to that title if you want to make it a bit more prolific [13:36] jiboumans: alrighty ... just be warned, they're going to look rather desktopy (putting a GTK frontend on a command line too) [13:36] jiboumans: right, already includes gsoc too [13:36] jiboumans: https://blueprints.edge.launchpad.net/testdrive/+spec/server-maverick-testdrive-frontend-gsoc [13:37] kirkland, see on line 194 ot f that pastebin. that to me is an error in euca [13:37] as euca-describe-images showed 'running' for an instance that had 0.0.0.0 as its IP [13:37] smoser: no, it didn't ... i manually terminated that one [13:38] smoser: i looked at your code and didn't see any way to exit out of wait_for_running() [13:38] i think it times out at 600 seconds [13:38] maybe "terminated" isn't terminated enough [13:39] i'm wrong. it never does [13:39] kirkland, you were right. [13:39] smoser: i only see one break in your while loop in wait_for_running() [13:39] ttx: i like what smoser has here, though ... [13:40] ttx: i think you/me/mathiaz/smoser should run some over night stress tests for the next few weeks [13:40] ttx: firing up random numbers of images, different machine sizes, and different images in combination [13:40] smoser: your script doesn't wait for termination ? [13:40] ttx: my cloud goes into powersave mode most nights [13:41] ttx: as i'd expect most of our should [13:41] ttx: for the next few weeks, i suggest that we run some random stuff overnight [13:41] ttx, it does not check state of instances to make sure they're terminated after terminate-instances [13:41] smoser: it can take some time to go from shutting-down to terminated, in my experience [13:41] but it does/did try 4 times with 10 second sleeps in between on the 'start' [13:42] so if a start returns failure it will just sleep and try again [13:42] ttx: btw, i'm doing a UEC demo at the Texas Linux Fest this saturday [13:42] ttx: do you have slides or the video from your similar presentation? [13:42] kirkland: right, you wanted to get my slides [13:42] (obviously that 40 second total should probably be increased) [13:42] * ttx uploads [13:43] smoser: i just set rounds to 1,000,000, and I'm just going to let it run until i need my cloud and ctrl-c it [13:43] smoser: it's tee'ing to a logfile [13:43] kirkland: http://one.ubuntu.com/p/Vy/ [13:43] you should probably redirect stderr to log also [13:43] * ttx tries ubuntuone magic [13:43] smoser: what are the indications of problems? error? failure? what should i grep for? [13:43] oooh, look at ttx with his ubuntu one url. fancy [13:44] "failed reach" is the real failure [13:44] ttx: thanks [13:44] ttx: i'm going to do a dry run practice today/tomorrow [13:44] it was already on ubuntuone, so I figured I'd try the "publish with ubuntuone" feature [13:44] there is a problem with phpldapadmin on lucid, is this the right place to report it? [13:44] working as designed :) [13:44] kirkland, ttx we should probably dump a bit more into that script to make it more reliable while not necissarily more forgiving. [13:44] a.) wait for terminate [13:44] macno: no please report it on launchpad [13:45] b.) wait longer for try-to-start [13:45] kirkland: are you going to have internet access ? [13:45] c.) maybe put warnings/errors somewhere other than stdout [13:45] d.) put a limit on wait-for-run [13:45] kirkland: the ukuug presentation did not have that access, and the installer went into strange network timeout quirks [13:47] * ttx stares at the oracle logo in the openoffice splash screen... [13:48] ttx: i'm not planning on having internet access [13:48] ttx: i'm planning on bring 2 of the dell 1200's, plus the linksys router [13:48] ttx: the router would connect and network the 2 laptops [13:49] ttx: if there is internet access, i can use the router to bridge to the wireless [13:49] ttx: if not, no big deal, i'm not going to fight with it [13:49] ttx: i'm also bringing my laptop, and i was planning on bringing a webcam too, that I could point at the other two machines while they install [13:49] ttx: and 2 usb keys [13:49] ttx: i'll do the install from usb key [13:49] ttx: rather than show my preseed magic [13:50] kirkland: just test beforehand how the installer reacts to network-but-no-internet situation... I had to press Cancel at a few file downloads at critical times during the install demo [13:50] ttx: right, the apt one? [13:50] kirkland: yes, doing it from the USB key tells better how easy it is for everyone to try out [13:51] kirkland: yes. Also the timeserver takes a long time to timeout [13:51] ttx: definitely === hggdh_ is now known as hggdh [14:06] someone knows if it's possible to configure dhcpd so it provides an ip range based on mac address? [14:09] i have an old pc that i want to use as a server....its a AMD Athlon runin at 1.73 GHz and 512 ram [14:09] the hd is 30 gb tho [14:13] oh no, mercurial 1.5 didn't make it into lucid [14:13] sigh [14:14] masu3701: thank you for that info [14:18] pmatulis: ? === dendrobates is now known as dendro-afk [14:24] ttx: I'm going the i386 server upgrade right now. [14:24] ScottK: ok. zul is running it as well [14:25] It's the server my IRC runs through, so if I disappear and don't come back, you'll know it didn't go well. [14:33] ttx: upgrade passed [14:46] New bug: #558427 in eucalyptus (main) "UEC should be a superset of the Ubuntu Server" [High,In progress] https://launchpad.net/bugs/558427 [14:51] ttx: Mine passed too. [14:51] one test to go before 100% coverage ! [15:16] ttx: I tried like described in the link posted by you to testcase UEC. The instance keeps pending and won't change to run state :/ [15:21] smoser is the rig availabel? === luis__lopez is now known as luis_lopez [15:28] ttx: can you add this one to your list https://bugs.edge.launchpad.net/ubuntu/+source/apache2/+bug/392759 [15:28] Launchpad bug 392759 in apache2 "[FFE] apache2 DoS attack using slowloris" [Unknown,Fix released] [15:32] zul: ack [15:32] ttx: thanks [15:34] Only 5 bugs filed. Not a bad upgrade. [15:35]