[01:24] hi SergioMeneses how goes life? [01:26] phillw, arriving at home and you? [01:26] I saw your email to ubuntu-laptop-testing [01:27] I'm doing well, shortly to bed ( 02:27 here) [01:31] phillw, that sounds awesome! I'm sleepy [01:33] I'm hopeful one will reply before I have to submit the time table to the classroom team and advertising. [01:33] you already pinged one people in there [01:34] s/people/person [01:36] phillw, perfect [01:39] SergioMeneses: if one of that team can say that will be the classroom sessions held in that time scale, I'm happy to let it stay as TBA [01:40] phillw, prime or carla should response soon ;) dont worry [01:40] I dont know if I can be present, so I dont take the responsibility yet [01:43] SergioMeneses: 36 hours, hopefully one / both will respond by then. It'd be a shame to lose it from our week of quality on the classroom area with the advertising. Heck, I got a new lubuntu tester by doing a 30 min session on ubuntu open week :D [01:44] phillw, ;) [01:49] phillw, I'm trying to get all isos for testing this weekend === _salem is now known as salem_ [03:13] SergioMeneses: great, soz, been busy with other little jobs..... It's been one of those days :) [03:13] phillw, like all us [03:17] SergioMeneses: yeah, I had a trainee bug triager needing a bit of help. Heck, we were all noobs once :) [03:19] TLoT was available to help reverse his couple of errors.... It happens to us all :) [03:53] Good morning [03:56] phillw, you're right [03:56] pitti, morning [03:57] hey SergioMeneses, how are you? [03:57] pitti, good :D [03:57] I'm goint out to bed now [03:57] guys always is good to talk to you, see you later [03:57] good night [04:02] good night! [04:17] hi pitti what you doing up this time of day? [04:17] fixing pm-utils and answering email :) [04:24] pitti: one day, I successfully get you to run a classroom session.... I've known you far too long for you to plead "I'm a N00b" :D [04:24] heh [04:24] I did a few in the past, but it's been a while [04:26] pitti: As a certain Mr Murray is doing one, I think that you should make yourself available :) [04:29] pitti: I'm even now looking into how we do this for the start of 14.04, as that will be an LTS, our classroom sessions in 13.04 were beta release, our 13.10 ones are RC. :) === salem_ is now known as _salem [05:10] I'm just finishing moving to 12.04, now that it is stable enough [05:10] and talk of 14.04 :) [05:34] Patrickdk: whilst we test ubuntu+1, we also have to plan for ubuntu+2 :) [05:35] ya, since it changed from testing to quality, I got completely lost [05:35] but managed to get my own personal 12.04 tested and rolled out only a few months behind the offical 12.04, just takes time to schedule things [05:36] we run at the bleeding edge, especially for testing tools :) [05:37] ya, I used to do the testing runs, for iscsi, jeos, and the like [05:38] mainly cause I'm concerned with ubuntu server [05:38] Patrickdk: oh, buggeration... that has reminded me that I need to set up the iSCSI server area..... [05:40] I just need to check that the ipV4 address is now free for me to re-use (They're getting as rare as hens' teeth :D ) [05:40] this is what nat -> nat -> nat -> nat -> nat is for :) [05:40] just don't attempt udp [05:41] they did just assign a /10 cgn block, that range is going get rather interesting [05:44] Patrickdk: I have a dedi server with 2 blocks of RIPE vpV4 addresses (4 in each), they are getting pretty hard to get hold of now, unless you want to pay a 'management' fee for your server :( [05:45] ya, I don't have to deal with that [05:46] have blocks from arin, so just have to pay the yearly $100 [05:46] and my servers are happy [05:47] Patrickdk: I pay 1GB / month for each of mine. I'm quite okay with that :D [05:48] I have as many ipV6's as I want (well, millions) for free, but ipV4 dedicated addresses (not shared) are getting rarer [05:49] s/1 GB / 1GBP [05:49] ya, I have been thinking I should get another /24 [05:50] i have enough to what number of VM's my little dedi server can cope with. [05:50] :) [05:50] I have almost 300 vm's on a single server [05:50] 16GB of RAM does run out :) [05:50] 144gigs ram there [05:51] no vm has <4gigs ram [05:51] I guess you use KVM for it? [05:51] no, vmware [05:51] kvm mem dedup just isn't good enough [05:51] why not the kernel ops? [05:52] and I can't remember if it got memory compression [05:52] kernel ops? [05:52] kvm is built into the kernel, it is the same as used for cloud [05:53] I know what kvm, and it still has serious limitations and slowness [05:53] if you don't exceed your memory, you will be fine with kvm [05:54] I still has some issues with kvm's paravirt drivers [05:54] but when I ran the same workload in kvm, and it took 45min, and ran it in vmware and it took 25min [05:54] I think I know what one I should use [05:56] indeed, just as the 'open cloud' players say it is open source, and then tie in their own 'private' stuff to make you pay for it :) [05:57] But, I could also choose Virtual Box.... such is life :) [05:57] oh, please no [05:57] I couldn't even get that to run stable enough to test with [05:58] in this case I was comparing for laptop usage, kvm vs vmware workstations vs virtualbox [05:58] Ahh, it argues the hell with the in built kernel kvm stuff unless you remove that part of the kernel 1st :D [05:59] I've got someone else to run the VBox classroom session, I'm sticking with the totally 100% free KVM :) [05:59] well, it's 3hours past bed, and critical san failure is fixed :) [06:00] sad when you have a full HA san unit [06:01] and before you cause a failure, you attempt to move load off the unit to be worked on, and it goes down hard [06:01] Patrickdk: good to chat, do keep in touch. I'm interested to know just how GPL vmware actually is. [06:01] well, that is the trick to vmware, like most other things, like redhat [06:01] it's very gpl [06:01] My dedi lives away from home in a server farm, less worries about a/c :) [06:01] except for the completely inhouse part :) [06:02] I'm trained in RedHat (even passed the exams). My dedi server runs CentOs, just to keep costs down. [06:03] 0 centos installs here [06:03] about 40 rhel6 installs [06:03] and around 40 or so ubuntu [06:03] oddly enough, all of mine are ubuntu server 12.04 LTS. funny how things go :) [06:03] and near 600 windows [06:04] No Scientific Linux users? [06:04] na [06:04] some suse, but only vmware stuff [06:04] I'm seriously considering moving to it. [06:05] CentOs people are pretty arrogant [06:05] all 2 of them? [06:05] lol, there are a lot of CentOs users :P [06:06] ya, but users don't roll out security updates [06:06] and they where lagging a good month+ behind rhel [06:06] why I would never touch them [06:06] all security stuff comes from rhek [06:06] *rhel [06:06] yes, and if it really takes a month to do a straight repack [06:07] and for rhel, you have to pay each month... hardly F/OSS :) [06:07] I rarely see rhel and ubuntu lagging a week [06:07] I don't mind paying [06:08] that is the least of my issues [06:08] I do, as my VM's are free to the people who use them. It is not a business for me. [06:21] good morning === mmrazik is now known as mmrazik|afk === forestpi1kie is now known as forestpiskie === forestpiskie is now known as Guest33673 === Guest33673 is now known as hobgoblin === mmrazik|afk is now known as mmrazik [07:41] $ run-adt-test -s -P deja-dup [07:41] tr: extra operand 'o' [07:41] quoi ? [07:42] ah, --proposed works; I'll look at that [08:18] jibel: did you ever see the "ssh ... localhost sudo adt-run .." hang forever in run-adt-test? [08:18] jibel: when I run that command standalone on a shell, it doesn't hang, but if I run it with the full "run-adt-test -sUkd squid3" it does (I verified that adt-run has finished in the VM) [08:20] * pitti goes to slam some -vv and strace around it [08:21] pitti, I didn't see this hang [08:21] jibel: ok, thanks; I'll debug [08:21] pitti, is it only with squid3 ? [08:21] yes, apparently === mmrazik is now known as mmrazik|ofp === mmrazik|ofp is now known as mmrazik [09:03] jibel: so, http://paste.ubuntu.com/5738202/ helps [09:04] jibel: I'm not really sure why, but the other ssh command there also uses it (forcing tty allocation) [09:12] pitti, but why would it fail only for squid3 [09:12] jibel: I don't know really; especially since running the same ssh command from a terminal works just fine [09:13] perhaps the squid3 test forks something, or does something to its stdout/err [09:13] our script does exec 2>&1, I'll try that with a manual ssh again [09:15] oh, interesting [09:16] so putting that "ssh localhost run-adt" command into a shell script still works; adding exec 2>&1 above it also still works [09:16] not sure what's really different from calling it from run-adt-test [10:20] who looks after the boot speed testing jenkins test ... [10:20] i am looking to understand how the machines are provisioned, and whether they [10:20] recently changes from raring to saucy [10:26] gema, ^^ [10:26] gema, specifically i am interested in the jump in the kernel grpahs here: http://reports.qa.ubuntu.com/bootspeed/machine/1/amd64/ [10:26] does that correspond to raring->saucy [10:28] apw: looking [10:31] apw: the results from 20130424 is the last raring image [10:31] the next one is 20130502, and that was already raring [10:32] apw: not sure what jump you refer to [10:32] apw: I am going to ask the guys to color code the bars if that is possible, to be able to see when we changed release [10:34] or add the release name to the runs table: http://reports.qa.ubuntu.com/bootspeed/job-bugs/image/ [10:35] gema, second and third bars have distict jumps in them about 7-8 bars from the right [10:35] gema, on this graph: http://reports.qa.ubuntu.com/bootspeed/machine/1/amd64/ [10:40] apw: the kernel graph or the generic one? [10:41] oh, you are talking about the kernel one [10:41] very clear jump [10:41] apw: do you know you can zoom in, right? [10:42] there is a jump up between the 20130520 and 20130524 [10:42] I remember we changed from server images to desktop images [10:42] but I don't remember exactly when [10:42] so I am going to find that out [10:43] in kernel init there's also a regression between the same two [10:44] apw: nuclearbob is not online right now [10:44] I can check with him tomorrow morning [10:44] he is in japan this month === mmrazik is now known as mmrazik|lunch === mmrazik|lunch is now known as mmrazik === hobgoblin is now known as forestpiskie === _salem is now known as salem_ [13:45] jibel: ping, got a moment to talk about otto? [14:37] shadeslayer, hey, sure, how can I help? [14:39] jibel: hiya, my primary questions are, a) Can we use otto to do upgrade testing like auto-upgrade-tester , b) Can otto be run on EC2 ( I don't think those have a graphics card ) and c) I can't seem to run otto on Kubuntu ISOs, screen goes blank, I read this is an issue and tried the commands in doc/README , still didn't work [14:39] s/we/I/ [14:40] unfortuantely, if I run otto on my work machine it kills my running KDE instance as well [14:41] shadeslayer, it'll probably need some fixes/updates to run with KDE, I only tried with lightdm, for example the autologin bits are really specific to this dm [14:42] jibel: Kubuntu uses lightdm :) [14:42] shadeslayer, ah ok, that shows my knowledge of kde :) I'll give it a try an see how it goes [14:43] well kubuntu [14:43] okay, thanks alot [14:43] what about a and b? :P [14:43] shadeslayer, regarding ec2, it should work as long as LXC in VMs is supported [14:44] isn't EC2 a LXC container itself? [14:44] because one can launch X on an EC2, but when I ran otto, it said that it couldn't find a graphics card [14:44] so I thought that maybe that's why it failed [14:45] shadeslayer, for a) you can do upgrade testing but really the purpose was to do UI testing with direct access to the hardware, so for upgrade testing you'd rather use lxc directly [14:45] okay === salem_ is now known as _salem === _salem is now known as salem_ === mmrazik is now known as mmrazik|afk === josepht_ is now known as josepht === mmrazik|afk is now known as mmrazik [19:28] balloons: joining us for this autopilot planning call? [19:29] thomi, sure thing [19:29] balloons: https://plus.google.com/hangouts/_/b73c6b041872d401efb62cef6a957823dd4bca5a [19:47] hey everyone :-) === fginther is now known as fginther|afk [20:23] hey DanChapman [20:23] busy day today :-) [20:25] balloons, Yeah been getting stuck into ubiquity took me a while to figure out a good host-to-vm solution then xorg kept crashing, grrr can't wait for that bug fix :D How's your day been? [20:26] good to see we can enhance terminal test :) [20:27] just met with the autopilot guys.. robotfuel has a way he thinks to work around the issue with gtk apps not introspecting [20:28] oh cool that sounds good.... when is that going to planned for? [20:28] he has to investigate it.. perhaps he'll share a bit more :-) [20:30] nice be good to get into firefox, thunderbird and maybe even have a group effort at libreoffice :D [20:31] balloons, Is the next hackfest going to be on the core ubuntu touch apps? [20:36] DanChapman, indeed we've invitied everyone who wishes to come [20:36] however, it's not like we can't do other things :-) [20:36] manual and other autopilot test hacking is always welcome :) [20:37] Sweet, I think i would like to carry on getting through the Gtk apps though. Be awesome to see everyone all hacking autopilot together across the board :D [20:38] exactly :-) See if we can get some new blood in [20:38] that sounds like such a weird term [20:38] ugh.. new blood [20:38] lol [20:38] how's ubiquity coming along? [20:41] yeah not bad been battling with my macbook though and xorg issues, am going to buy a thinkpad E530 over the weekend and get rid of this thing. Then can get into it properly hopefully without the issues i am getting now :-( [20:42] mm.. sorry to hear of the issues [20:44] I will keep battling on for now though. :D [20:45] :-) [20:45] At least ive learnt my lesson ;-) and won't lose my test again [20:46] lol.. indeed! [20:47] this was the solution i ended up with http://pastebin.ubuntu.com/5739965/ once setup it easier than using bzr branches === salem_ is now known as _salem [20:53] balloons: it doesn't look like there is a way for nautilus to spawn a new process like gnome-terminal, but you can use a subprocess of nautilus -q (which makes nautlius quit) before launching it with autopilot [20:54] robotfuel, so what we be the full steps needed to get it working? [20:55] DanChapman, nice set of notes [20:55] I wonder if we can share that somehow [20:55] balloons: I'll pastebin something [20:55] thanks :-) [21:00] balloons: http://pastebin.ubuntu.com/5740039/ [21:01] robotfuel, ahh.. I was just curious about your other idea [21:01] but indeed it works [21:02] autopilot vis now sees nautilus and the new version launched is happy to be introspected :-) [21:02] ty! [21:03] balloons: the non-hack way to fix this would be to write a autopilot gtk module that lets autopilot find the new child process that is launched. [21:03] balloons: I don't have experience with writing gtk modules though. [21:04] balloons: it would be kind of like the at-spi bridge for the accessibility layer [21:05] robotfuel,indeed letting autopilot find child processes would solve several issues for us [21:05] balloons: ap already does search child processes [21:05] the problem here s that the child process does not load the introspection interface [21:06] thomi, I knew you were going to say that.. lol, we talked about it with ubiquity :-) [21:06] balloons: there's even tests in autopilot to cnfirm that :) [21:06] ohh! nice! does this happen on the QT side also, or just with gtk? === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha