--- Log opened do sep 18 00:00:25 2014 00:03 <@danken> apuimedo: beyond starting to test el7? 00:03 <@danken> (that was the reason we saw that firewalld issue) 00:04 < apuimedo> danken: well, I mean, the el7 policy should have been close to the one in f19-f20 00:04 < apuimedo> which we tested a lot 00:04 < apuimedo> I'll have to ask mgrepl where they took the policy from 00:04 < apuimedo> ... 00:05 < nsoffer> danken, similar issue on f20 00:05 < nsoffer> sanlock is not allowed to send SIGTERM to vdsm 00:05 < nsoffer> thats why we see 40 retires in sanlock log 00:05 < nsoffer> but on f20, sanlock is allowed to send SIGKILL to vdsm 00:06 -!- fromani: has quit [Quit: Leaving] 00:06 < apuimedo> nsoffer: so it is not allowed in f19 and el7 00:06 < apuimedo> but it is on f20? 00:07 < nsoffer> I did not test this on 19 yet 00:07 < apuimedo> ok 00:12 -!- rmatinata (Ricardo Marin Matinata): has joined #vdsm 00:16 -!- adahms (Andrew Dahms): has joined #vdsm 00:23 -!- pkliczew (Piotr Kliczewski): has joined #vdsm 00:24 -!- apuimedo: has quit [Ping timeout: 260 seconds] 00:25 <@danken> apuimedo: few upstream users have shops based on fedora. I think 95% of our install base is el6. 00:25 -!- danken: has quit [Quit: Leaving.] 00:30 < nsoffer> ok, on f19, it is the same as rehl7 00:30 < nsoffer> we could catch this bug more than a year ago 00:30 -!- pkliczew: has quit [Ping timeout: 255 seconds] 00:48 -!- #vdsm dougsland: has quit [Ping timeout: 245 seconds] 01:26 -!- adahms_ (Andrew Dahms): has joined #vdsm 01:28 -!- adahms: has quit [Ping timeout: 272 seconds] 01:30 -!- adahms_ is now known as adahms-mtg 01:42 -!- nsoffer: has quit [Ping timeout: 272 seconds] 02:02 -!- adahms-mtg: has quit [Ping timeout: 272 seconds] 02:14 -!- adahms-mtg (Andrew Dahms): has joined #vdsm 02:34 -!- dougsland (Douglas): has joined #vdsm 02:52 -!- adahms-mtg: has quit [Ping timeout: 246 seconds] 03:08 -!- adahms-mtg (Andrew Dahms): has joined #vdsm 03:17 -!- adahms-mtg: has quit [Ping timeout: 246 seconds] 03:30 -!- adahms-mtg (Andrew Dahms): has joined #vdsm 03:53 -!- mliu (purple): has joined #vdsm 03:53 -!- mliu: has left #vdsm [] 04:09 -!- wgao: has quit [Ping timeout: 245 seconds] 04:38 -!- wgao (wgao): has joined #vdsm 04:41 -!- adahms-mtg: has quit [Ping timeout: 272 seconds] 04:48 -!- gpadgett: has quit [Quit: Leaving] 04:53 -!- wgao: has quit [Ping timeout: 255 seconds] 04:55 -!- wgao (wgao): has joined #vdsm 04:57 -!- adahms-mtg (Andrew Dahms): has joined #vdsm 05:45 -!- vered: has quit [Quit: Leaving] 06:29 -!- fabiand (Fabian Deutsch): has joined #vdsm 06:54 -!- bazulay: has quit [Quit: Leaving.] 07:01 -!- bazulay (purple): has joined #vdsm 07:11 -!- aravindavk (Aravinda): has joined #vdsm 07:38 -!- ibarkan (ibarkan): has joined #vdsm 07:54 -!- sbonazzo (purple): has joined #vdsm 07:59 -!- fromani (Francesco Romani): has joined #vdsm 08:01 -!- fpliger (fpliger): has joined #vdsm 08:14 -!- mliu: has quit [Client Quit] 08:16 -!- timothy (Timothy Asir): has joined #vdsm 08:16 -!- timothy is now known as Guest42811 08:20 -!- adahms-mtg: has quit [Ping timeout: 272 seconds] 08:21 -!- ndarshan (Darshan n): has joined #vdsm 08:30 -!- pkliczew (Piotr Kliczewski): has joined #vdsm 08:32 -!- adahms-mtg (Andrew Dahms): has joined #vdsm 08:47 -!- phoracek (phoracek): has joined #vdsm 08:50 -!- xaviern (Xavier): has joined #vdsm 08:50 -!- apuimedo (antoni): has joined #vdsm 09:01 -!- bazulay: has quit [Ping timeout: 240 seconds] 09:06 -!- ibarkan: has quit [Ping timeout: 245 seconds] 09:08 -!- #vdsm xaviern: has quit [Ping timeout: 260 seconds] 09:18 -!- mpolednik (Martin Polednik): has joined #vdsm 09:20 -!- ibarkan (ibarkan): has joined #vdsm 09:21 -!- acanan (Aharon Canan): has joined #vdsm 09:23 -!- phoracek: has quit [Quit: WeeChat 0.4.3] 09:23 -!- phoracek (phoracek): has joined #vdsm 09:23 -!- adahms-mtg: has quit [Quit: Leaving] 09:29 -!- xaviern (Xavier): has joined #vdsm 09:30 -!- rmatinata: has quit [Ping timeout: 246 seconds] 09:32 -!- nsoffer (Nir Soffer): has joined #vdsm 09:34 -!- mode/#vdsm: by ChanServ 09:34 -!- danken (purple): has joined #vdsm 09:35 -!- shaohef (shaohef): has joined #vdsm 09:35 -!- shaohef: has left #vdsm [] 09:38 < nsoffer> ybronhei, can you take http://gerrit.ovirt.org/33020 ? 09:39 -!- wgao: has quit [Ping timeout: 260 seconds] 09:40 < nsoffer> ybronhei, that was quick :-) 09:41 < ybronhei> nsoffer: wasn't so quick, i saw that before 09:42 -!- phoracek: has quit [Quit: WeeChat 0.4.3] 09:43 -!- rmatinata (Ricardo Marin Matinata): has joined #vdsm 09:44 -!- acanan: has quit [Quit: Leaving] 09:44 -!- acanan (Aharon Canan): has joined #vdsm 09:44 < msivak_> nsoffer: no, I actually wrote the commit message, but forgot to save it.. yesterday was too busy.. 09:45 < nsoffer> msivak_, I guessed 09:45 < nsoffer> msivak_, about the "x$blh" - not needed - right? 09:45 < nsoffer> msivak_, hold habits? 09:45 < nsoffer> old :-) 09:45 < msivak_> well.. it used to be that it was necessary 09:46 < nsoffer> msivak_, when you do not quote I think 09:46 < nsoffer> x$blah = foo 09:46 < msivak_> you had to quote anyways because of spaces 09:46 < msivak_> there was an issue with empty arguments to test iirc 09:46 < nsoffer> with quotes? I never seen this 09:48 < msivak_> http://stackoverflow.com/questions/174119/why-do-shell-script-comparisons-often-use-xvar-xyes see the accepted answer, the second case (option processing) 09:48 -!- wgao (wgao): has joined #vdsm 09:51 < msivak_> nsoffer: I was taught this ages ago by http://tldp.org/LDP/abs/html/comparison-ops.html#FTN.AEN3669 09:51 < msivak_> nsoffer: so no, it is probably not needed here, but I would not call that old idiom 09:52 < msivak_> now.. where my commit message went... 09:52 -!- ibarkan: has quit [Ping timeout: 258 seconds] 09:52 -!- ibarkan_ (ibarkan): has joined #vdsm 09:53 -!- #vdsm ybronhei: has quit [Ping timeout: 246 seconds] 09:53 -!- bazulay (purple): has joined #vdsm 09:56 < nsoffer> msivak_, I guess that: is safe, even if working with broken shell/test 09:57 < msivak_> yep, it just feels weird, I always write the variable first 09:58 -!- #vdsm xaviern: has quit [Quit: WeeChat 0.4.3] 09:58 < nsoffer> msivak_, you started this paranoia :-) 09:59 -!- xaviern (Xavier): has joined #vdsm 09:59 < msivak_> nsoffer: it is not about broken shell, this was a standard defensive way of writing shell scripts.. I do not think about using it, it is automatic for me 10:00 < msivak_> and the option processing case is still valid even with modern shells 10:00 < nsoffer> msivak_, ok, call it broken world, where you cannot write code that make sense 10:00 < msivak_> well that is bash :) 10:00 < msivak_> we are lucky to deal with linux (and bash) only 10:01 < nsoffer> well if this script run on ubuntu, you may get dash 10:02 < msivak_> true, not sure what kinks they have 10:02 < msivak_> but I suppose it should be pretty much compatible 10:02 -!- mliu: has left #vdsm [] 10:04 < nsoffer> msivak_, all the hooks run as filters? 10:04 < msivak_> nsoffer: no, just this one 10:04 < msivak_> nsoffer: well two, but the restore is not implemented yet 10:04 < nsoffer> msivak_, it is much nicer like this 10:05 < nsoffer> then vdsm can even stack them 10:05 < msivak_> nsoffer: there is one small issue with libvirt hooks 10:05 < msivak_> nsoffer: they support only one file with exact name 10:06 < nsoffer> we can change that to hooktype.d/foo 10:06 -!- ibarkan_: has quit [Quit: Konversation terminated!] 10:07 < msivak_> nsoffer: well if we need more hooks then yes, we would have to write our dispatcher 10:07 -!- ibarkan_ (ibarkan): has joined #vdsm 10:09 -!- ybronhei (purple): has joined #vdsm 10:09 < fromani> apuimedo: hi 10:09 < apuimedo> fromani: hi 10:10 < fromani> apuimedo: I'm having a memory lapse... does libnl need to access special files on /proc or /sys or any other virtual filesystem? 10:11 < apuimedo> fromani: no, I don't think so 10:11 < apuimedo> it just opens a netlink socket with the kernel 10:11 < apuimedo> fromani: why? 10:12 < fromani> apuimedo: just curious, trying to pinpoint ppc build errors on the mock env 10:12 < fromani> trying to help pinpoint, actually 10:12 < apuimedo> fromani: is this related to the mail about python-ethtool? 10:12 < nsoffer> msivak_, if the hook is a filter, when the event is not relevant, should we return 0? 10:13 < nsoffer> the code check for empty output? 10:13 < fromani> apuimedo: well, my train of thought started from there, yes 10:14 < nsoffer> msivak_, the hook api is not clear - seems that we should return the input unchanged 10:14 < nsoffer> e.g. if:; then exec cat; fi 10:14 < apuimedo> fromani: that python-ethtool version is older than the one in which david sommerseth and I made fixes to allow a process to have different libs open netlink sockets without conflict 10:15 < fromani> apuimedo: so it is *very* old :) 10:15 < apuimedo> well, a bit old 10:15 < apuimedo> but not ancient 10:15 < apuimedo> xd 10:15 < fromani> :) 10:15 < msivak_> nsoffer: the documentation says that if we return anything else then 0 it will be logged 10:15 < msivak_> nsoffer: right at the bottom of http://www.libvirt.org/hooks.html 10:16 < msivak_> nsoffer: I do not think we want that 10:16 < apuimedo> fromani: danken says that the python-ethtool release version for the platform is 0.11, so that one is fine 10:16 -!- bazulay: has quit [Ping timeout: 272 seconds] 10:16 < apuimedo> iirc 10:16 < fromani> apuimedo: ok then 10:16 < msivak_> nsoffer: and migration is the only filter there is with restore coming 10:16 < fromani> apuimedo: just trying to learn/remember something :) 10:16 < msivak_> nsoffer: but I see what you mean 10:17 < msivak_> nsoffer: but they deal with it gracefully: Empty output is identical to copying the input XML without changing it. 10:17 < nsoffer> msivak_, ok 10:19 < nsoffer> msivak_, can you comment about this is the script? I'm sure someone else will be confused by this 10:24 < msivak_> sure 10:25 -!- fpliger: has quit [Remote host closed the connection] 10:27 -!- fpliger (fpliger): has joined #vdsm 10:32 -!- bazulay (purple): has joined #vdsm 11:14 -!- phoracek (phoracek): has joined #vdsm 11:18 -!- phoracek: has quit [Ping timeout: 245 seconds] 11:29 < msivak_> fromani: ping, do you still have the migration setup? 11:30 < fromani> msivak_: yep, I do, but I'm chasing another strange issue apparently involving json-rpc 11:30 < fromani> msivak_: is this afternoon good enough> 11:30 < fromani> ? 11:30 < msivak_> I suppose so 11:30 < fromani> msivak_: ok, will prioritize this task as much as I can 11:31 * fromani too many things rolling... 11:31 < msivak_> I will not be on irc between about 1 and 4 11:31 < fromani> msivak_: hopefully I'll have the patch V+1 by 16:00 11:31 < msivak_> thanks 11:31 < fromani> np 11:31 < msivak_> migration almost works for me now :) 11:32 < msivak_> uh.. lost connection with qemu process again.. 11:33 < msivak_> there is something weird going on indeed 11:33 -!- mpolednik: has quit [Ping timeout: 272 seconds] 11:34 -!- moolit (purple): has joined #vdsm 11:34 -!- phoracek (phoracek): has joined #vdsm 11:34 < msivak_> guys is there anything with F19, oVirt 3.4.4 and latest libvirt/qemu there that would prevent migrations? 11:35 < msivak_> do we have any contacts in the qemu team? I already tried libvirt, but all looks good there 11:36 < fromani> ok, I got fooled by the lack of libvirtd logs... we reduced the verbosiness, right. Moving on... 11:36 * fromani feels dumb 11:36 -!- moolit: has quit [Read error: Connection reset by peer] 11:36 -!- moolit (purple): has joined #vdsm 11:50 -!- moolit: has quit [Quit: Leaving.] 11:56 -!- adahms (Andrew Dahms): has joined #vdsm 12:23 -!- ndarshan: has quit [Quit: Leaving] 12:29 -!- ibarkan_: has quit [Quit: Konversation terminated!] 12:39 -!- derez_ (Daniel Erez): has joined #vdsm 12:40 -!- ibarkan (ibarkan): has joined #vdsm 12:41 < apuimedo> pkliczew: ping 12:52 -!- bazulay: has quit [Quit: Leaving.] 12:53 -!- bazulay (purple): has joined #vdsm 12:59 -!- bazulay: has quit [Ping timeout: 245 seconds] 13:00 -!- shaohef (shaohef): has joined #vdsm 13:08 -!- pkliczew: has quit [Remote host closed the connection] 13:10 -!- pkliczew (Piotr Kliczewski): has joined #vdsm 13:10 < pkliczew> apuimedo, pong 13:11 < apuimedo> pkliczew: for the bug you reported 13:11 < apuimedo> how was it first configured? 13:11 < apuimedo> you answered the needinfo 13:11 < apuimedo> saying that after the problem you edited the ifcfg 13:11 < apuimedo> but how was the net set up before that? 13:11 < apuimedo> NM? 13:11 < pkliczew> apuimedo, I haven't checked, I only can say that it was default config with dhcp 13:12 < pkliczew> apuimedo, I installed fresh centos6 vm and run vdsm on it (nested) 13:13 < apuimedo> pkliczew: mmm 13:13 < pkliczew> apuimedo, no changes was done to network interface and I haven't looked at it before failure 13:13 < apuimedo> it's strange that dhcp was not passed to the setupNetworks command 13:13 < apuimedo> due to that, it lost connectivity and had to roll back 13:13 < pkliczew> apuimedo, the vm was running on a host with vdsm so we had vdsm in vdsm 13:13 < apuimedo> that's fine 13:14 < pkliczew> apahim, roll back is ok but why the nic was removed? 13:14 -!- bazulay (purple): has joined #vdsm 13:15 < apuimedo> pkliczew: because it didn't have a previous vdsm configuration and thus 13:15 < apuimedo> it rolled back to nothing 13:15 < apuimedo> (that is the major bug, btw) 13:16 < pkliczew> apuimedo, there was nic for sure I sshed to this vm and lost connection after roll back 13:16 < apuimedo> it couldn't read the original config 13:16 < apuimedo> hence it didn't know how to return to it 13:16 < apuimedo> why it couldn't read it? 13:16 < apuimedo> I have to find out 13:16 < pkliczew> I would love to give you more details but I haven't checked net config 13:18 < apuimedo> pkliczew: I'll see what I can do 13:18 < apuimedo> thanks 13:18 < pkliczew> thank you! 13:23 < apuimedo> no, thank you for the report ;-) 13:23 -!- #vdsm dougsland: has quit [Ping timeout: 245 seconds] 13:35 -!- aravindavk: has quit [Ping timeout: 272 seconds] 13:36 -!- phoracek: has quit [Quit: WeeChat 0.4.3] 13:51 -!- aravindavk (Aravinda): has joined #vdsm 13:56 -!- moolit (purple): has joined #vdsm 14:05 -!- aravindavk: has quit [Ping timeout: 260 seconds] 14:16 -!- aravindavk (Aravinda): has joined #vdsm 14:40 < fromani> apuimedo: out of curiosity, did you ever upgraded a box from F19 to F20? I mean without reinstalling from scratch 14:40 -!- shaohef: has quit [Quit: Leaving.] 14:41 < apuimedo> fromani: 5-10 times 14:41 < fromani> apuimedo: did you followed an HOWTO or something? 14:41 < fromani> follow* 14:41 < apuimedo> nope 14:41 < apuimedo> yum --releasever=20 update 14:41 < fromani> apuimedo: cool, and that worked OK? 14:42 < apuimedo> fromani: most of the time 14:42 < fromani> apuimedo: no nasty surprises? I do have a F19 box which is dear to me (= a pain to backup and restore) 14:42 < fromani> apuimedo: ok, cool 14:42 < apuimedo> well, but your home partition is safe 14:42 < apuimedo> isn't it? 14:43 < apuimedo> fromani: the recommended method is using fedup 14:43 < apuimedo> but it never worked too well for me 14:43 < fromani> apuimedo: ok, got it 14:43 < fromani> apuimedo: sure the home is safe (and will be backuped just in case...) 14:44 < apuimedo> fromani: so then go ahead ;-) 14:44 < apuimedo> the worst that can happen is to bootstrap a new system from a usb flash running gentoo or arch 14:44 < apuimedo> ;-) 14:45 < fromani> apuimedo: I guess yes, I'll clean things up before... the less package to upgrade, the less chance to blow up 14:45 < apuimedo> fromani: indeed 14:54 -!- aravindavk: has quit [Ping timeout: 260 seconds] 15:00 < fromani> msivak_: tested migration succesfully. Since restore cannot be fixed through hook, I ticked V+1. Documentation about verification is on gerrit. 15:00 -!- nsoffer: has quit [Read error: Connection reset by peer] 15:01 < msivak_> fromani: thanks a lot 15:01 < fromani> msivak_: you're welcome 15:01 -!- aravindavk (Aravinda): has joined #vdsm 15:03 -!- gpadgett (Greg Padgett): has joined #vdsm 15:04 -!- fpliger_ (fpliger): has joined #vdsm 15:06 -!- fpliger: has quit [Ping timeout: 246 seconds] 15:08 -!- rmatinata: has quit [Ping timeout: 272 seconds] 15:10 -!- apuimedo is now known as apuimedo|lunch 15:11 -!- nsoffer (Nir Soffer): has joined #vdsm 15:12 -!- dyasny (Dan Yasny): has joined #vdsm 15:12 -!- #vdsm dyasny: has quit [Read error: Connection reset by peer] 15:13 -!- shaohef (shaohef): has joined #vdsm 15:13 -!- dyasny (Dan Yasny): has joined #vdsm 15:16 -!- nsoffer: has quit [Read error: Connection reset by peer] 15:19 -!- Guest42811: has quit [Ping timeout: 245 seconds] 15:29 -!- adahms: has quit [Quit: Leaving] 15:39 -!- apuimedo|lunch is now known as apuimedo 15:52 -!- rmatinata (Ricardo Marin Matinata): has joined #vdsm 15:55 -!- shaohef: has quit [Quit: Leaving.] 15:57 -!- aravindavk: has quit [Ping timeout: 258 seconds] 16:03 -!- derez_: has quit [Quit: Leaving] 16:04 -!- derez_ (Daniel Erez): has joined #vdsm 16:05 -!- timothy_ (Timothy Asir): has joined #vdsm 16:06 < fromani> danken: I'm pleased to mark http://gerrit.ovirt.org/#/c/27026/4 as Verified 16:07 < fromani> it took (much) longer than expected for funny incidental reasons but at last is verified... 16:11 -!- timothy_: has quit [Ping timeout: 240 seconds] 16:28 -!- #vdsm xaviern: has quit [Ping timeout: 272 seconds] 16:30 < fromani> thanks for the merge danken ! 16:33 <@danken> fromani: no, thank you. 16:35 < fromani> danken: the more I think to sampling, the more your idea to have a separate process becomes appealing 16:48 < apuimedo> danken: I have a darned bug 16:48 < apuimedo> exposed by Zordrak 16:52 < apuimedo> I assumed in my ip -4 addr reporting 16:52 < apuimedo> sorry 16:52 < apuimedo> on my 'addr' reporting in netinfo 16:52 < apuimedo> that any secondary address would have the secondary flag 16:52 < apuimedo> however, in el7 and F20 16:52 < apuimedo> you can have multiple primary addresses per interface 16:53 < apuimedo> and I'm returning the last one 16:53 < apuimedo> I'm thinking that I should change this to return the first one that the kernel reports 16:53 < apuimedo> (it's the one used by the engine for the .vv file) 16:55 -!- orc_emac: has quit [Quit: leaving] 16:55 -!- ibarkan: has quit [Ping timeout: 245 seconds] 16:55 -!- xaviern (Xavier): has joined #vdsm 17:00 -!- vered: has quit [Ping timeout: 272 seconds] 17:04 -!- phoracek (phoracek): has joined #vdsm 17:06 -!- acanan: has quit [Ping timeout: 240 seconds] 17:11 <@danken> apuimedo: ah, that's the reason for his inconsistent ADDR0 GATEWAY1 ? 17:12 < apuimedo> very related, yes 17:12 * apuimedo reading the kernel code 17:12 <@danken> apuimedo: trusting the kernel order does not sound very stable 17:12 < apuimedo> they kept the possibility of setting the secondary flag 17:12 < apuimedo> but it is optional now 17:12 < apuimedo> danken: well 17:13 < apuimedo> the better idea is to give the addr that has the default route 17:13 < apuimedo> if it is in that device 17:13 -!- #vdsm xaviern: has quit [Ping timeout: 240 seconds] 17:13 < apuimedo> (which it may not be) 17:13 < apuimedo> and look at the source routing otherwise) 17:13 < apuimedo> gets ugly 17:13 -!- derez_: has quit [Quit: Leaving] 17:13 -!- pkliczew: has quit [Ping timeout: 255 seconds] 17:15 < sbonazzo> guys, quick question 17:15 < sbonazzo> virsh dumpxml --domain 7f9cd40e-252e-4bed-8247-5e2221fe9807 17:15 < sbonazzo> Please enter your authentication name: 17:17 < apuimedo> sbonazzo: already answered ;-) 17:17 < sbonazzo> apuimedo: thanks 17:18 < apuimedo> you're welcome 17:20 <@danken> sbonazzo: better use `virsh -r` than say "shibboleth" 17:20 < apuimedo> or just do what I do 17:21 < apuimedo> configure libvirt auth in /etc 17:21 < apuimedo> and then forget about it ;-) 17:22 < fromani> danken: msivak_ wait a moment, just noticed a build failure on jenkins 17:23 -!- aravindavk (Aravinda): has joined #vdsm 17:23 < fromani> danken: msivak_ http://jenkins.ovirt.org/job/vdsm_master_create-rpms_merged/1871/ and http://jenkins.ovirt.org/job/vdsm_master_create-rpms-el6-x86_64_merged/5/ 17:23 < fromani> danken: msivak_ on F20 and centos7 rpm builds OK 17:23 * fromani looking what jenkins complains for... 17:24 < fromani> it's just centos6, this makes sense 17:24 < fromani> http://jenkins.ovirt.org/job/vdsm_master_create-rpms_merged/label=centos6/ 17:28 <@danken> fromani: it compained about disk sapce; or was there something real? 17:28 < fromani> danken: I'm concerned about this 17:28 < fromani> RPM build errors: 17:28 < fromani> File not found: /home/jenkins/workspace/vdsm_3.4_install_rpm_sanity_gerrit/label/centos6/rpmbuild/BUILDROOT/vdsm-4.14.16-1.git552955c.el6.x86_64/etc/libvirt/hooks/qemu 17:28 < fromani> make: ***: Error 1 17:29 -!- phoracek: has quit [Ping timeout: 245 seconds] 17:30 < fromani> danken: msivak_ probably another if in the package file list is enough 17:30 * fromani really needs to read 'maximum rpm' again... 17:30 < sbonazzo> danken: thanks 17:31 <@danken> fromani: could you point me to that log? 17:32 < fromani> danken: of course: http://jenkins.ovirt.org/job/vdsm_3.4_install_rpm_sanity_gerrit/70/ 17:33 < fromani> danken: is in the backport for branch ovirt-3.4 17:33 < fromani> danken: from here I got suspects 17:33 <@danken> fromani: ah of course 17:33 <@danken> msivak_: since we install the hook only on Fedora 17:34 <@danken> msivak_: we must not list it in %files of el6 17:34 <@danken> please send a follow-up fix for that? 17:34 -!- aravindavk: has quit [Quit: Leaving] 17:35 <@danken> fromani: could you give a quick -1 for the pending stable patches? 17:35 < fromani> danken: msivak_ sorry for noticing this so late :( 17:35 < fromani> danken: already gave 17:35 <@danken> fromani: better your late than my never 17:37 -!- sbonazzo: has quit [Quit: Leaving.] 17:38 -!- phoracek (phoracek): has joined #vdsm 18:10 -!- fpliger_: has quit [Remote host closed the connection] 18:11 -!- fpliger (fpliger): has joined #vdsm 18:24 -!- fromani: has quit [Quit: Leaving] 18:26 -!- fpliger: has quit [Remote host closed the connection] 18:28 -!- fpliger (fpliger): has joined #vdsm 18:35 -!- apuimedo: has quit [Ping timeout: 272 seconds] 18:38 -!- bazulay: has quit [Ping timeout: 260 seconds] 18:41 -!- fpliger: has quit [Remote host closed the connection] 18:41 -!- fpliger (fpliger): has joined #vdsm 18:46 -!- phoracek: has quit [Quit: WeeChat 0.4.3] 18:57 -!- pkliczew (Piotr Kliczewski): has joined #vdsm 19:05 -!- pkliczew: has quit [Ping timeout: 245 seconds] 19:18 -!- fpliger_ (fpliger): has joined #vdsm 19:21 -!- fpliger: has quit [Ping timeout: 272 seconds] 19:30 -!- timothy_ (Timothy Asir): has joined #vdsm 19:31 -!- phoracek (phoracek): has joined #vdsm 19:33 -!- fpliger_: has quit [Remote host closed the connection] 19:33 -!- fpliger (fpliger): has joined #vdsm 19:35 -!- timothy_: has quit [Ping timeout: 272 seconds] 19:35 -!- moolit: has quit [Quit: Leaving.] 19:35 -!- timothy_ (Timothy Asir): has joined #vdsm 19:35 -!- phoracek: has quit [Client Quit] 19:36 -!- phoracek (phoracek): has joined #vdsm 19:38 -!- fpliger: has quit [Ping timeout: 272 seconds] 20:04 -!- fpliger (fpliger): has joined #vdsm 20:09 -!- fpliger: has quit [Ping timeout: 272 seconds] 20:45 -!- nsoffer (Nir Soffer): has joined #vdsm 20:46 -!- apahim: has quit [Ping timeout: 272 seconds] 20:54 -!- timothy_: has quit [Ping timeout: 246 seconds] 20:56 -!- apahim (Amador Pahim): has joined #vdsm 21:05 -!- #vdsm ybronhei: has quit [Quit: Leaving.] 21:05 -!- fpliger (fpliger): has joined #vdsm 21:08 -!- xaviern (Xavier): has joined #vdsm 21:09 -!- fpliger: has quit [Ping timeout: 260 seconds] 21:13 -!- #vdsm xaviern: has quit [Ping timeout: 255 seconds] 21:18 -!- fpliger (fpliger): has joined #vdsm 21:50 < nsoffer> msivak_, ... lost connection with qemu - happen to me all the time on fedora 20 - on what system do you get this? 22:01 -!- apuimedo (antoni): has joined #vdsm 22:02 < nsoffer> alitke__, ping 22:02 < alitke__> nsoffer, pong 22:02 < nsoffer> alitke__, on what platforms are you testing? 22:03 < nsoffer> on the hosts 22:03 < alitke__> fedora 20 mostly. RHEL7 here and there 22:04 < nsoffer> and did you notice any issues on fedora 20? 22:04 < nsoffer> e.g. live snapshot? 22:04 < nsoffer> live storage migration? 22:04 < nsoffer> vm migration? 22:04 < nsoffer> all seems to be broken on my fedora 20 hosts 22:05 -!- shaohef (shaohef): has joined #vdsm 22:05 < alitke__> live snapshots are broken with jsonrpc 22:05 < nsoffer> I tried with xmlrpc 22:05 < alitke__> haven't tested storage or vm migration much 22:05 < nsoffer> unless there is a bug where you select xmlrpc and it actually use jsonrpc 22:05 -!- fpliger: has quit [Remote host closed the connection] 22:06 < alitke__> make sure that jsonrpc didn't corrupt your storage domain. 22:06 < nsoffer> do you know what cause the issue with jsonrpc and live snap? 22:06 < alitke__> you might be using V1 when you think you're using v3 22:06 -!- fpliger (fpliger): has joined #vdsm 22:06 < alitke__> not yet. 22:06 < alitke__> fromani and I have opened a few bugs where we're seeing strange issues. 22:07 < nsoffer> I don't think it is this - I create new storage domains or used very old ones, created few month ago before jsonrpc could corrupt them :-) 22:07 < nsoffer> alitke__, are you using real hosts or nested kvm? 22:08 < alitke__> I have done both. 22:08 < nsoffer> my are currently fedora 20 on bare metal, running fake hosts on virt-manager 22:08 < nsoffer> maybe my trouble is related to fedora 20 upgrade last week? 22:09 < nsoffer> before that I don't recall any issues, where it run on fedora 19 22:10 < alitke__> Are you using virt-preview or regular f20 repo for libvirt and qemu? 22:10 < nsoffer> I think virt-preview 22:10 < alitke__> for the record, the issues seem to be recent for me. 22:10 < alitke__> ok... If you are not trying live merge it would be interesting to downgrade to fedora main repo and see if the problems persist. 22:11 -!- fpliger: has quit [Ping timeout: 272 seconds] 22:11 < nsoffer> do you have fedora 20 machine up now? 22:12 < nsoffer> can you try live snapshot? 100% failing here 22:12 < nsoffer> start a vm, create snapshot, vm pause and die 22:12 < alitke__> saw this earlier too... Let me try 22:12 < nsoffer> "lost connection to qemu process..." in vdsm log 22:13 < nsoffer> btw, I see also selinux errors for qemu, trying to access tun device 22:15 < alitke__> Yeah, this sounds like what I was also seeing. 22:16 < nsoffer> for example: 22:16 < nsoffer> type=AVC msg=audit(1410971197.122:88026): avc: denied { read } for pid=20278 comm="qemu-system-x86" path="/dev/net/tun" dev="devtmpfs" ino=11278 scontext=system_u:system_r:svirt_t:s0:c26,c368 tcontext=system_u:object_r:svirt_image_t:s0:c515,c850 tclass=chr_file permissive=0 22:16 < nsoffer> type=SYSCALL msg=audit(1410971197.122:88026): arch=c000003e syscall=0 success=no exit=-13 a0=1a a1=7fd1a8d6031c a2=11000 a3=0 items=0 ppid=1 pid=20278 auid=4294967295 uid=107 gid=107 euid=107 suid=107 fsuid=107 egid=107 sgid=107 fsgid=107 tty=(none) ses=4294967295 comm="qemu-system-x86" exe="/usr/bin/qemu-system-x86_64" subj=system_u:system_r:svirt_t:s0:c26,c368 key=(null) 22:17 < alitke__> ok, snapshot worked on host using xmlrpc 22:17 -!- fabiand: has quit [Ping timeout: 260 seconds] 22:17 < nsoffer> real host? 22:18 < alitke__> yes 22:18 < nsoffer> what libvirt/qemu version? 22:18 -!- apuimedo: has quit [Ping timeout: 272 seconds] 22:18 < alitke__> I notice there is a libvirt update I don't have installed... let me look 22:19 < nsoffer> I update both the host and the nested kvm hosts yesterday 22:19 < alitke__> $ sudo yum info libvirt-daemon-kvm 22:19 < alitke__> Loaded plugins: langpacks, refresh-packagekit 22:19 < alitke__> Installed Packages 22:19 < alitke__> Name : libvirt-daemon-kvm 22:19 < alitke__> Arch : x86_64 22:19 < alitke__> Version : 1.2.8 22:19 < nsoffer> so running with latest bugs 22:19 < alitke__> Release : 2.fc20 22:19 < alitke__> Size : 0.0 22:19 < alitke__> Repo : installed 22:19 < alitke__> From repo : fedora-virt-preview 22:19 < alitke__> Summary : Server side daemon & driver required to run KVM guests 22:19 < alitke__> URL : http://libvirt.org/ 22:19 < alitke__> License : LGPLv2+ 22:19 < alitke__> Description : Server side daemon and driver required to manage the 22:19 < alitke__> : virtualization capabilities of the KVM hypervisor 22:19 < alitke__> Available Packages 22:19 < alitke__> Name : libvirt-daemon-kvm 22:19 < alitke__> Arch : x86_64 22:19 < alitke__> Version : 1.2.8 22:19 < alitke__> Release : 3.fc20 22:19 < alitke__> Size : 35 k 22:19 < alitke__> Repo : fedora-virt-preview/20/x86_64 22:19 < alitke__> Summary : Server side daemon & driver required to run KVM guests 22:19 < alitke__> URL : http://libvirt.org/ 22:19 < alitke__> License : LGPLv2+ 22:19 < alitke__> Description : Server side daemon and driver required to manage the 22:19 < alitke__> : virtualization capabilities of the KVM hypervisor 22:20 < alitke__> oops... That was supposed to be an fpaste 22:20 < alitke__> let me try to update 22:21 < nsoffer> I have: # rpm -q libvirt-daemon-kvm 22:21 < nsoffer> libvirt-daemon-kvm-1.2.8-2.fc20.x86_64 22:21 -!- apuimedo (antoni): has joined #vdsm 22:21 -!- shaohef: has quit [Quit: Leaving.] 22:21 < nsoffer> so same version as yours 22:22 -!- fpliger (fpliger): has joined #vdsm 22:22 < nsoffer> I think need to check again with real host 22:33 < alitke__> nsoffer, Yeah, I am creating snapshots fine (both hosts using xmlrpc 22:34 < nsoffer> ok, then it is probably env issue on my side 22:34 < alitke__> But I 22:34 < alitke__> I've seen this on other systems so not so sure we should dismiss it so easilt. 22:34 < alitke__> easily 22:37 < nsoffer> alitke__, do you remember what kind of systems? 22:38 < alitke__> I believe it was one of the tlv test hosts 22:38 < alitke__> are those nested virt? 22:38 < nsoffer> maybe 22:43 -!- apuimedo: has quit [Ping timeout: 245 seconds] 23:14 -!- moolit (purple): has joined #vdsm 23:37 -!- apahim: has quit [Quit: Leaving] --- Log closed vr sep 19 00:00:27 2014