--- Log opened Tue Feb 18 00:00:19 2014 00:11 -!- #vdsm fsimonce: has quit [Quit: Coyote finally caught me] 00:15 -!- danken_afk: has quit [Quit: Leaving.] 01:43 -!- moolit: has left #vdsm [] 02:04 -!- ybronhei: has quit [Ping timeout: 245 seconds] 02:27 -!- apuimedo: has quit [Ping timeout: 272 seconds] 02:35 -!- abaron_: has quit [Ping timeout: 245 seconds] 02:35 -!- abaron: has quit [Ping timeout: 260 seconds] 03:05 -!- shaohef (shaohef): has joined #vdsm 03:07 -!- bala (purple): has joined #vdsm 03:22 -!- zhshzhou (Zhou Zheng Sheng): has joined #vdsm 03:32 -!- bala: has quit [Quit: Leaving.] 03:57 -!- acathrow (Aic): has joined #vdsm 03:57 -!- acathrow: has left #vdsm [] 04:06 -!- acathrow (Aic): has joined #vdsm 04:41 -!- sahina (Sahina Bose): has joined #vdsm 04:42 -!- sahina: has quit [Remote host closed the connection] 04:58 -!- mohankumar__ (M. Mohan Kumar): has joined #vdsm 05:05 -!- mohankumar__: has quit [Ping timeout: 245 seconds] 05:06 -!- mohankumar__ (M. Mohan Kumar): has joined #vdsm 05:12 -!- mohankumar__: has quit [Ping timeout: 265 seconds] 05:13 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:18 -!- mohankumar: has quit [Ping timeout: 265 seconds] 05:21 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:25 -!- mohankumar: has quit [Read error: Connection reset by peer] 05:26 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:31 -!- mohankumar: has quit [Ping timeout: 260 seconds] 05:31 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:36 -!- vered (Vered Volansky): has joined #vdsm 05:36 -!- mohankumar: has quit [Ping timeout: 260 seconds] 05:37 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:38 -!- dougsland (Douglas): has joined #vdsm 05:42 -!- mohankumar: has quit [Ping timeout: 260 seconds] 05:44 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 05:50 -!- mohankumar: has quit [Ping timeout: 260 seconds] 06:14 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 06:14 -!- timothy (Timothy Asir): has joined #vdsm 06:14 -!- ybronhei (purple): has joined #vdsm 06:15 -!- bala (purple): has joined #vdsm 06:20 -!- mohankumar: has quit [Ping timeout: 248 seconds] 06:21 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 06:27 -!- mohankumar: has quit [Read error: Connection reset by peer] 06:31 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 07:07 -!- mode/#vdsm: by ChanServ 07:07 -!- danken (purple): has joined #vdsm 07:08 -!- abaron (Ayal Baron): has joined #vdsm 07:08 -!- abaron_ (Ayal Baron): has joined #vdsm 07:24 -!- timothy: has quit [Ping timeout: 245 seconds] 07:28 -!- aravindavk (Aravinda): has joined #vdsm 07:33 -!- mohankumar: has quit [Ping timeout: 248 seconds] 07:35 -!- sbonazzo (purple): has joined #vdsm 07:44 -!- danken: has quit [Ping timeout: 260 seconds] 07:55 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 07:55 -!- danken (purple): has joined #vdsm 07:55 -!- mode/#vdsm: by ChanServ 07:56 -!- abaron_: has quit [Ping timeout: 265 seconds] 07:57 -!- abaron: has quit [Ping timeout: 265 seconds] 08:01 -!- fromani (Francesco Romani): has joined #vdsm 08:08 -!- nsoffer (Nir Soffer): has joined #vdsm 08:20 -!- ewarszaw (Eduardo Warszawski): has joined #vdsm 08:24 -!- pkliczew_ (Piotr Kliczewski): has joined #vdsm 08:30 -!- mohankumar: has quit [Ping timeout: 240 seconds] 08:36 -!- ybronhei: has quit [Quit: Leaving.] 08:37 -!- ybronhei (purple): has joined #vdsm 08:40 -!- shaohef: has quit [Ping timeout: 265 seconds] 08:43 -!- xaviern (Xavier): has joined #vdsm 08:47 -!- shaohef (shaohef): has joined #vdsm 08:55 -!- #vdsm,#theforeman xaviern: has quit [Ping timeout: 260 seconds] 09:00 -!- fabiand (Fabian Deutsch): has joined #vdsm 09:03 -!- timothy (Timothy Asir): has joined #vdsm 09:09 -!- abaron_ (Ayal Baron): has joined #vdsm 09:10 -!- abaron (Ayal Baron): has joined #vdsm 09:12 -!- amuller (Assaf Muller): has joined #vdsm 09:24 -!- fromani_ (Francesco Romani): has joined #vdsm 09:25 -!- fromani_: has quit [Client Quit] 09:26 -!- fromani_ (Francesco Romani): has joined #vdsm 09:27 -!- fromani: has quit [Ping timeout: 264 seconds] 09:27 -!- acanan_ (Aharon Canan): has joined #vdsm 09:28 -!- fromani_ is now known as fromani 09:34 -!- xaviern (Xavier): has joined #vdsm 09:41 -!- ewarszaw: has quit [Ping timeout: 265 seconds] 09:45 -!- fsimonce (Federico): has joined #vdsm 09:50 -!- vered: has quit [Ping timeout: 252 seconds] 09:51 -!- zhshzhou: has quit [Remote host closed the connection] 09:57 -!- acanan_: has quit [Quit: Leaving] 09:58 -!- acanan (Aharon Canan): has joined #vdsm 10:05 -!- vered (Vered Volansky): has joined #vdsm 10:22 -!- nsoffer: has quit [Ping timeout: 260 seconds] 10:24 -!- apuimedo (antoni): has joined #vdsm 10:26 -!- ewarszaw_ (Eduardo Warszawski): has joined #vdsm 10:27 -!- vered: has quit [Ping timeout: 272 seconds] 10:31 -!- saggi (purple): has joined #vdsm 10:31 -!- acanan: has quit [Ping timeout: 248 seconds] 10:37 -!- osvoboda (Ondřej Svoboda): has joined #vdsm 10:39 < osvoboda> danken: Hangout in 20 minutes? I will call you. 10:39 -!- vered (Vered Volansky): has joined #vdsm 10:41 -!- osvoboda: has left #vdsm [] 10:46 -!- acanan (Aharon Canan): has joined #vdsm 10:48 -!- osvoboda (Ondřej Svoboda): has joined #vdsm 11:03 -!- shaohef: has quit [Ping timeout: 240 seconds] 11:04 -!- ykaplan (Yeela Kaplan): has joined #vdsm 11:04 -!- acanan: has quit [Ping timeout: 260 seconds] 11:07 < apuimedo> danken: https://bugzilla.redhat.com/show_bug.cgi?id=1015009 11:16 -!- osvoboda: has quit [Quit: osvoboda] 11:16 -!- acanan (Aharon Canan): has joined #vdsm 11:20 -!- acanan: has quit [Client Quit] 11:20 -!- acanan (Aharon Canan): has joined #vdsm 11:21 -!- timothy: has quit [Ping timeout: 248 seconds] 11:24 -!- nsoffer (Nir Soffer): has joined #vdsm 11:28 -!- osvoboda (Ondřej Svoboda): has joined #vdsm 11:32 -!- apuimedo: has quit [Ping timeout: 245 seconds] 11:38 -!- timothy (Timothy Asir): has joined #vdsm 11:46 -!- apuimedo (antoni): has joined #vdsm 11:49 -!- timothy: has quit [Read error: Operation timed out] 11:50 -!- amuller: has quit [Ping timeout: 252 seconds] 11:52 < nsoffer> fsimonce, ping 11:52 < fsimonce> nsoffer, pong 11:53 -!- moolit (purple): has joined #vdsm 11:57 -!- apahim (Amador Pahim): has joined #vdsm 11:59 -!- mohankumar (M. Mohan Kumar): has joined #vdsm 11:59 -!- mohankumar: has quit [Client Quit] 12:00 -!- amuller (Assaf Muller): has joined #vdsm 12:01 -!- timothy (Timothy Asir): has joined #vdsm 12:04 -!- hchiramm_ (Humble Chirammal): has joined #vdsm 12:08 < apuimedo> danken: let me know if you get to reproduce the "connection was reset by peer" 12:10 -!- amuller: has quit [Read error: Operation timed out] 12:23 -!- saggi: has quit [Ping timeout: 272 seconds] 12:30 -!- aravindavk: has quit [Ping timeout: 248 seconds] 12:31 <@danken> apuimedo: /me is too busy doing non-programming. but I would, if I see it. 12:36 -!- amuller (Assaf Muller): has joined #vdsm 12:36 -!- timothy: has quit [Read error: Operation timed out] 12:39 < amuller> apuimedo: Any ideas why the network functional tests aren't running on Gerrit? 12:40 < apuimedo> amuller: because at some point it was crashing and the ci team disabled them 12:40 < apuimedo> :P 12:40 < amuller> ok 12:40 < apuimedo> amuller: when and if I have some time 12:40 < apuimedo> I have to enable them again and make sure that it doesn't take the jenkins slave offline 12:42 -!- fromani: has quit [Quit: Leaving] 12:43 -!- danken: has quit [Quit: Leaving.] 12:44 -!- mode/#vdsm: by ChanServ 12:44 -!- danken (purple): has joined #vdsm 12:46 -!- fromani (Francesco Romani): has joined #vdsm 12:49 -!- timothy (Timothy Asir): has joined #vdsm 12:54 -!- saggi (purple): has joined #vdsm 12:59 -!- saggi: has quit [Ping timeout: 252 seconds] 12:59 < nsoffer> danken, check http://gerrit.ovirt.org/24636 and http://gerrit.ovirt.org/24614 13:00 < nsoffer> apuimedo, your pool turned out nice and simple 13:00 < nsoffer> apuimedo, but I think we don't need the lazy population 13:00 < apuimedo> nsoffer: thanks 13:01 < apuimedo> nsoffer: danken suggested creating the socket when it's needed to do away with the population port 13:01 < apuimedo> *part 13:01 -!- saggi (purple): has joined #vdsm 13:01 < apuimedo> and making the pool a singleton that is accessed by a method that will create the instance if needed 13:01 < nsoffer> apuimedo, you mean lazy creation, stop when pool is full? 13:01 < apuimedo> nsoffer: yes 13:02 < nsoffer> that is even harder, I like simple solutions 13:02 < apuimedo> I like lazy stuff as well, but I'm tempted by the simplicity 13:02 < apuimedo> :P 13:02 -!- bala: has quit [Ping timeout: 260 seconds] 13:02 < nsoffer> anyone is going to complain if we create 3 nl sockets on startup? 13:03 < apuimedo> nsoffer: even creating the maximum amount is negligible in time 13:03 < apuimedo> less than half second for sure 13:03 < apuimedo> nsoffer: there is a reason I called the _nl_handle_destroy how I did 13:03 < apuimedo> it is a different opeartion from closing the socket 13:03 < nsoffer> half a second to create 3 sockets?! 13:03 < apuimedo> no, 1024 13:03 < apuimedo> it takes less than that 13:03 < nsoffer> ok 13:04 < nsoffer> apuimedo, I'm sure you did the right thing, but the api looks confusing 13:04 < nsoffer> _get_nl_socket looks like "get me one of your sockets" 13:04 < nsoffer> but it creates a new one 13:04 < apuimedo> nsoffer: I'm just hesitant on calling it close, since it could give the idea on first read that I'm just closing 13:05 < nsoffer> and _nl_handle_destroy is a lower level lbnl function 13:05 < apuimedo> and missing freeing the handle (which frees the port form libnl's internal map( 13:05 < apuimedo> ) 13:05 < apuimedo> ) 13:05 < apuimedo> yes 13:05 < nsoffer> I don't understand the difference 13:05 < apuimedo> maybe I could say 13:05 < apuimedo> create socket 13:05 < apuimedo> sorry 13:05 < apuimedo> _socket_create 13:05 < apuimedo> and _socket_destroy 13:05 < apuimedo> for simmetry 13:06 < apuimedo> the difference is that nl_close is not freeing the handle 13:06 < nsoffer> to make it clear that this is not libnl api, I would use xxx_socket 13:06 < apuimedo> and the resources 13:06 < nsoffer> libnl -> _nl_xxx_xxx 13:06 < apuimedo> ok 13:06 < nsoffer> module -> xxx_socket 13:07 -!- ykaplan: has quit [Ping timeout: 272 seconds] 13:08 < nsoffer> apuimedo, but what happen if we have 3 network operations that take all the sockets 13:08 < nsoffer> and then sampling have no available socket? 13:08 < nsoffer> is this possible? 13:08 < apuimedo> it could be, getVdsCaps is not under any lock 13:08 -!- ykaplan (Yeela Kaplan): has joined #vdsm 13:09 < nsoffer> maybe sampling can use its own socket? 13:09 < apuimedo> so if one were to do N non-serial operations that used sockets and the pool was size N-1 there'd be a wait 13:09 < apuimedo> however 13:10 < apuimedo> getVdsCaps and sampling are the only non-serialized access that I can see 13:10 < apuimedo> nsoffer: I thought about that 13:10 < nsoffer> I guess that sampling should never wait 13:10 < apuimedo> socket ownership 13:10 < nsoffer> but request from the engine can wait 13:10 < apuimedo> I'd rather it wouldn't yes 13:11 < nsoffer> maybe we don't even need a pool 13:11 < nsoffer> just create one socket for sampling and another for requests from engine 13:12 < nsoffer> put a lock around the socket so multiple callers will wait 13:13 < nsoffer> and go a longer to walk with your baby :-) 13:13 < apuimedo> heh 13:13 < apuimedo> I'm not sure it is that easy to tell which lock you should get 13:14 < nsoffer> You need a NLSocket object for that 13:14 < nsoffer> which will use a lock to serialize access 13:19 < apuimedo> danken: what do you think about this ^^ 13:19 < apuimedo> ? 13:23 -!- timothy: has quit [Ping timeout: 272 seconds] 13:27 -!- acathrow (Aic): has joined #vdsm 13:28 -!- acathrow: has left #vdsm [] 13:30 <@danken> apuimedo: since nl calls are short, a pool with size==1 seems like a good idea 13:30 < apuimedo> danken: what nsoffer proposes is that sampling will have its own socket 13:30 < apuimedo> IIUC 13:31 < apuimedo> and that the rest will contend for the single socket in netlink.py 13:32 < apuimedo> it should be very easy to do 13:37 * apuimedo is wrong 13:37 < apuimedo> not that easy 13:37 < apuimedo> netlink now lets you get a socket, but sampling.py uses ipwrapper, not netlink 13:37 < apuimedo> so I would have to increase the coupling 13:37 < apuimedo> :( 13:38 < nsoffer> apuimedo, give a socket to ipwarpper 13:38 < apuimedo> that is quite ugly 13:38 < nsoffer> I'm sure nobody will complain 13:38 < apuimedo> you ask ipwrapper.getLinks 13:38 < apuimedo> and you pass it a socket 13:39 < nsoffer> no, ipwrapper will keep its own socket 13:39 < apuimedo> it hurts my aestetic sense xD 13:39 < apuimedo> nsoffer: and how does ipwrapper tell when a request is for sampling and when not? 13:39 < nsoffer> apuimedo, I need to see the code so I can talk less bs :-) 13:39 < apuimedo> without coupling 13:39 < apuimedo> look at vdsm/sampling.py 13:39 < apuimedo> and lib/vdsm/ipwrapper.py:getLinks 13:44 < nsoffer> apuimedo, later 13:44 < apuimedo> ok 13:45 -!- AlbertoSilva (AlbertoSilva): has joined #vdsm 13:53 -!- moolit: has quit [Remote host closed the connection] 13:59 -!- ybronhei: has quit [Quit: Leaving.] 14:01 -!- ybronhei (purple): has joined #vdsm 14:02 -!- nsoffer: has quit [Ping timeout: 248 seconds] 14:02 -!- moolit (purple): has joined #vdsm 14:03 -!- amuller: has quit [Ping timeout: 252 seconds] 14:05 -!- vered: has quit [Ping timeout: 260 seconds] 14:10 <@danken> apuimedo: sorry, I failed to understand how nl is related or needed in ipwrapper 14:10 -!- amuller__ (Assaf Muller): has joined #vdsm 14:10 <@danken> as long as we use the ip command line, why do we care about nl there? 14:10 -!- amuller_: has quit [Read error: Operation timed out] 14:11 <@danken> amuller__: hi, is defroute now passed properly? which argument have we been overriding? 14:14 < apuimedo> danken: ipwrapper uses netlink.py for creating its Link objects 14:14 < apuimedo> which are then consumed from sampling.py and from netinfo 14:14 -!- saggi: has quit [Ping timeout: 272 seconds] 14:15 <@danken> apuimedo: /me is forgrtting code to quickly. 14:16 <@danken> what is the motivation for this? do you expect a real contention problem? 14:16 -!- dougsland (Douglas): has joined #vdsm 14:16 <@danken> I'd rather hide the socket inside netlink 14:21 < apuimedo> danken: well, I was trying not to delay the sampling 14:21 < apuimedo> but I'd just keep everything inside netlink 14:21 < apuimedo> and keep the pool :P 14:21 -!- ykaplan: has quit [Remote host closed the connection] 14:26 -!- ykaplan (Yeela Kaplan): has joined #vdsm 14:28 -!- saggi (purple): has joined #vdsm 14:29 <@danken> saggi: what are these repeated failures: http://jenkins.ovirt.org/job/vdsm_unit_tests_gerrit/7301/console 14:29 <@danken> FAIL: testLeakFd (miscTests.WatchCmd) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/jenkins/workspace/vdsm_unit_tests_gerrit/vdsm/tests/miscTests.py", line 988, in testLeakFd self.assertEquals(openFdNum(), openFds) AssertionError: 78 != 79 14:30 < dougsland> danken, ybronhei should I build 3.3.4 ? 14:31 < ybronhei> dougsland: why ? i saw you already built it yesterday 14:31 < dougsland> ybronhei, I have build from 3.4 branch not 3.3.x 14:34 < ybronhei> dougsland: let me check the mails 14:37 < dougsland> saggi, ping, pvt msg 14:37 < ybronhei> dougsland: what tag we have in our latest 3.3 build? 4.13.4 ? 14:37 < ybronhei> dougsland: can you check the latest patch there? 14:38 -!- dougsland (Douglas): has joined #vdsm 14:39 < ybronhei> dougsland: ok, last build was 4.13.3 14:39 < ybronhei> dougsland: the only commit that is missed there is vm iface statistics: never report negative rates 14:40 < ybronhei> dougsland: afaiu I think we should build 4.14.4 14:40 < dougsland> ybronhei, last build: spec: replace requires vdsm-python-cpopen 14:40 < dougsland> we are 11 patches behind. 14:40 < ybronhei> sorry, 4.13.3-4 14:40 < ybronhei> 11 ? how do you see that? 14:40 < dougsland> git log versus fedora spec 14:41 < ybronhei> ok, add those patches and create 4.13.3-4 14:41 < ybronhei> but just a moment, I check if something wait for review 14:42 < dougsland> ybronhei, ok, just let me know if you want me to do the build 14:42 < ybronhei> only http://gerrit.ovirt.org/#/q/status:open+project:vdsm+branch:ovirt-3.3,n,z which missing the validation 14:42 < ybronhei> dougsland: yes, please do the build 14:43 < dougsland> ybronhei, ok 14:43 <@danken> dougsland: ybronhei: there's a serious problem in our dropping of vdsm-python-cpopen: 14:43 <@danken> $ rpm -qp --obsoletes vdsm-python-cpopen-4.13.0-11.fc20.x86_64.rpm 14:43 <@danken> python-cpopen 14:43 < dougsland> ybronhei, IMHO these patches can wait. 14:44 <@danken> this means that we make yum THINK that vdsm-python-cpopen-4.13.0-11 replaces ANY kind of python-cpopen. 14:44 < ybronhei> danken: but we don't install it 14:45 < ybronhei> danken: unless it was installed before ? .. but what can we do with that 14:45 < dougsland> so far as I recall, we just build it. 14:45 < ybronhei> danken: its a buggy package . we can't change it backwards 14:45 -!- ewarszaw_: has quit [Quit: Konversation terminated!] 14:45 <@danken> ybronhei: but we MUST release a newer package to Fedora. 14:46 < ybronhei> of vdsm-python-cpopen ? don't we have 4.13.2 for that? 14:46 < ybronhei> we do 14:46 < ybronhei> vdsm-python-cpopen-4.13.2-1 14:47 < dougsland> I have marked as stable vdsm-4.13.3-3 14:47 < ybronhei> danken: ^ and this does not obsolete anything 14:48 <@danken> ybronhei: why isn't it on Fedora? 14:48 < ybronhei> danken: what about 4.13.3-4 ^^ do you suggest to build new release? 14:48 < ybronhei> it should... http://koji.fedoraproject.org/koji/buildinfo?buildID=484051 14:48 -!- shaohef (shaohef): has joined #vdsm 14:48 < dougsland> danken, it was in testing, just marked as stable 14:48 < ybronhei> send a link dougsland 14:49 < dougsland> https://admin.fedoraproject.org/updates/FEDORA-2014-1965/vdsm-4.13.3-3.fc20 14:49 < dougsland> https://admin.fedoraproject.org/updates/FEDORA-2014-1977/vdsm-4.13.3-3.fc19 14:49 <@danken> ybronhei: it is "obsoleted" https://admin.fedoraproject.org/updates/FEDORA-2013-23161/vdsm-4.13.2-1.fc20 14:50 < ybronhei> danken: i just checked - rpm -qp --obsoletes vdsm-python-cpopen-4.13.2-1.fc20.armv7hl.rpm , no output 14:50 <@danken> dougsland: ah, 4.13.3 is not yet pushed to stable 14:51 < ybronhei> but 4.13.2 also fine, we removed the obsolete. this should be in stable 14:51 <@danken> ybronhei: my problem is that 4.13.2 and 4.13.3 are not in stable 14:51 < dougsland> exactly 14:51 < dougsland> danken, I have requested, as soon koji updates 14:52 < dougsland> ok, going to build 3.3.4 14:54 <@danken> oh, 10 minutes ago. bodhi - 2014-02-18 13:46:40 14:54 <@danken> This update has been submitted for stable by dougsland. 14:56 -!- bala (purple): has joined #vdsm 14:57 -!- bala: has quit [Remote host closed the connection] 14:57 < dougsland> danken, yep. 15:00 -!- bala (purple): has joined #vdsm 15:02 -!- amuller__: has quit [Ping timeout: 264 seconds] 15:07 -!- amuller__ (Assaf Muller): has joined #vdsm 15:09 < amuller__> danken: yeah from my tests + functional tests it's ok now. We were overriding the third ipv6 argument 15:11 -!- bala1 (purple): has joined #vdsm 15:13 -!- bala: has quit [Ping timeout: 265 seconds] 15:14 <@danken> amuller__: but how we have missed that until now? 15:14 -!- amuller__ is now known as amuller 15:15 < amuller> danken: I have no idea, it's really weird 15:16 < amuller> danken: It could have surfaced by a user with more than one network (with a gateway), but that's not common / nobody does that in development 15:16 < amuller> as we never set defRoute, it was implicitly True for all networks 15:16 < amuller> So it behaved like in oVirt 3.2- 15:25 -!- dougsland (Douglas): has joined #vdsm 15:27 -!- rharper (Ryan Harper): has joined #vdsm 15:28 -!- nsoffer (Nir Soffer): has joined #vdsm 15:29 -!- osvoboda: has quit [Quit: osvoboda] 15:30 -!- osvoboda (Ondřej Svoboda): has joined #vdsm 15:46 -!- osvoboda: has quit [Ping timeout: 240 seconds] 15:55 < sbonazzo> dougsland: ybronhei : danken how is it going with 3.3.4 vdsm packages? 15:55 < dougsland> sbonazzo, working on it. 15:55 <@danken> 3.3.4? I'm still working on 3.4.0 beta 15:55 <@danken> beta 3 15:56 < sbonazzo> danken: 3.4.0 beta3 packages are already in the repo, douglas sent them yesterday 15:57 <@danken> sbonazzo: argh, then amuller's fix would have to wait for rc 15:59 < amuller> danken: Oh thanks for pushing it to ovirt-3.4 branch 15:59 < dougsland> danken, is that urgent fix? 16:00 < apuimedo> dougsland: well, for source routing it is quite important 16:01 -!- bala1: has quit [Quit: Leaving.] 16:01 < dougsland> if you guys want, I can build it after 3.3.4 16:02 < sbonazzo> guys, I don't really know what do you really think about upstream releases, but for me, any time I've to postpone by one day a release is a personal failure. 16:03 < dougsland> I meant, if we are still in time.. 16:05 -!- shaohef: has quit [Ping timeout: 252 seconds] 16:05 -!- ykaplan: has quit [Quit: Leaving] 16:06 -!- ykaplan (Yeela Kaplan): has joined #vdsm 16:06 < amuller> sbonazzo: dougsland: apuimedo: danken: This is a regression from 3.3 to 3.4 beta. It only affects users who have hosts that have more than a single network with a gateway defined. It worked fine in 3.3 and it's broken now in 3.4, like it was in 3.2. I'm not sure that this scenario happens often, or enough to justify the delay of a release. I'm personally fine with the fix going in the next scheduled build. 16:07 < amuller> I don't think it's important enough to delay a build or do something out of the ordinary, especially considering that we're talking about builds that are in the middle of a development cycle, before the first official release of 3.4. 16:08 -!- ykaplan: has quit [Client Quit] 16:09 -!- ykaplan (Yeela Kaplan): has joined #vdsm 16:12 < sbonazzo> amuller: thanks for the clarification 16:17 -!- sbonazzo: has quit [Ping timeout: 252 seconds] 16:18 -!- sbonazzo (purple): has joined #vdsm 16:27 < apuimedo> danken: I've embedded htb in drr 16:34 <@danken> apuimedo: and? everything explodes? 16:36 <@danken> dougsland: what did you use for 3.3.4? commit 79a0e17 ? 16:36 < apuimedo> danken: I'm about to test it :P 16:37 < apuimedo> I just thought rnetlink would give me the middle finger even before starting 16:37 < apuimedo> danken: but! On el6 I could create drr but on f20 I couldn't 16:37 < apuimedo> I'll have to check what's going on with fedora 16:38 <@danken> apuimedo: btw, master passes the whole test suit on my el6 16:38 < apuimedo> danken: ssl on or off? 16:38 <@danken> apuimedo: hmm 16:39 < apuimedo> http://fpaste.org/78256/37865139/ 16:39 <@danken> ssl=False 16:39 < apuimedo> aha! 16:39 <@danken> I'll try again with True 16:39 < apuimedo> :P 16:39 < apuimedo> thanks! 16:47 < dougsland> danken, this one and more 10 patches missing from our last build 16:53 -!- #vdsm,#theforeman xaviern: has quit [Ping timeout: 240 seconds] 17:01 -!- ykaplan: has quit [Ping timeout: 272 seconds] 17:05 -!- dougsland (Douglas): has joined #vdsm 17:06 -!- ykaplan (Yeela Kaplan): has joined #vdsm 17:11 -!- pkliczew_: has quit [Ping timeout: 245 seconds] 17:12 -!- saggi: has quit [Ping timeout: 265 seconds] 17:18 -!- xaviern (Xavier): has joined #vdsm 17:22 -!- AlbertoS_ (AlbertoSilva): has joined #vdsm 17:26 -!- ykaplan: has quit [Ping timeout: 272 seconds] 17:29 < fabiand> hey 17:30 < fabiand> can someone tell em how vdsm gathers cpu infos? 17:32 * fabiand found it 17:35 -!- amuller: has quit [Quit: Konversation terminated!] 17:40 -!- #vdsm,#theforeman xaviern: has quit [Ping timeout: 265 seconds] 17:43 -!- ykaplan (Yeela Kaplan): has joined #vdsm 17:49 -!- acanan: has quit [Ping timeout: 245 seconds] 17:50 <@danken> fabiand: good! 17:51 <@danken> apuimedo: I fail to use ssl=True altogether! crab. 17:52 <@danken> nsoffer: I did not debug this myself; have you seen issue with ssl=True on el6? 17:52 < nsoffer> no 17:54 < nsoffer> danken: I saw this http://jenkins.ovirt.org/job/vdsm_unit_tests_gerrit_el/6368/testReport/junit/jsonRpcTests/ReactorTests/test_rt__tcp___ssl_True_ 17:54 < nsoffer> danken, and the other errors I cannot remember the details 17:55 -!- apuimedo: has quit [Ping timeout: 260 seconds] 17:58 -!- ybronhei: has quit [Quit: Leaving.] 17:59 -!- ybronhei (purple): has joined #vdsm 17:59 < nsoffer> danken, I wonder how asyncore based server (using non-blocking io) is setting a timeout on a socket (making it blocking)? 18:03 <@danken> nsoffer: I'm refering to my own host, no json involved. 18:03 < nsoffer> danken, what issues? 18:03 <@danken> nsoffer: server blocks forever 18:04 < nsoffer> blocks how? 18:04 < nsoffer> does not respond to requests? 18:04 <@danken> nsoffer: I did not debug that, so I do not know when, but yes, it does not respond. 18:05 < nsoffer> the only thing I can think of, is failure to use vdsm with curl 18:05 < nsoffer> works with engine, when you try to send a request with curl (.e.g curl -X POST ... -d "data"), you get no response 18:06 < nsoffer> seen about 2 weeks ago 18:06 < nsoffer> And I used curl -k which works with ssl 18:06 < nsoffer> I also tried my own prototype using xmlrpcserver, and curl worked with it, but not with vdsm 18:07 < nsoffer> danken, I'll mail you the prototype 18:07 -!- fromani: has quit [Quit: Leaving] 18:09 -!- dougsland (Douglas): has joined #vdsm 18:11 < nsoffer> danken, check your mail 18:12 <@danken> nsoffer: thanks, but I'm not sure I could go over it :-( 18:13 < nsoffer> danken, what do you mean? 18:13 * danken has no time... 18:13 < nsoffer> its only 40 lines 18:13 < nsoffer> server.py 18:14 < nsoffer> danken, anyway, did you try to make request to vdsm using curl? 18:15 <@danken> nsoffer: yeah, and it blocks, but I do not really know what to make of this. 18:18 -!- sbonazzo: has quit [Quit: Leaving.] 18:24 < nsoffer> danken, I suspect our ssl socket code 18:24 < nsoffer> in my prototype, there is a one line hack: 18:24 <@danken> nsoffer: may well be so, but it has served us for years 18:24 < nsoffer> server.socket = ssl.wrap_socket(server.socket, certfile='server.pem', server_side=True) 18:25 -!- osvoboda (Ondřej Svoboda): has joined #vdsm 18:28 -!- saggi (purple): has joined #vdsm 18:30 < nsoffer> danken, we use some crypto library - why? 18:31 <@danken> you mean the m2crypto? 18:31 < nsoffer> yes 18:31 <@danken> it was reintroduced my Juan 18:31 <@danken> since it gives session caching 18:31 < nsoffer> used by xmlrpc? 18:32 <@danken> 5ea9b646f4fcd56ca48ed673810e86e26d44b523 18:33 < nsoffer> danken, interesting 18:33 < nsoffer> we can drop this if we merge http://gerrit.ovirt.org/24294 18:34 <@danken> because there are only few sessions 18:35 <@danken> btw, does the task.prepare cleanup pass some sanity tests? 18:35 < nsoffer> danken, I still did not have time to verify it in the current form 18:36 < nsoffer> and jenkins ci jobs fail with upstream code 18:36 < nsoffer> typically it fail to build the engine :-) 18:37 < nsoffer> danken, time to rewrite engine in python :-) 18:45 < nsoffer> danken, you fixed some thread safety issue there few month ago 18:52 -!- abaron_: has quit [Ping timeout: 260 seconds] 18:53 -!- abaron: has quit [Ping timeout: 272 seconds] 18:53 <@danken> nsoffer: yeah, m2crypto is not loved by me. 18:56 < nsoffer> danken, I don't have anything against it (I know nothing about it), but there is probably something wrong in secure xmlrpc server 18:57 < nsoffer> danken, it may be also curl bug, but It is unlikely 19:03 <@danken> nsoffer: indeed. but note that we use client certificates, which is pretty uncommon and I certainly did not enabled it in curl. 19:06 -!- nsoffer: has quit [Ping timeout: 260 seconds] 19:09 -!- saggi: has quit [Quit: Leaving.] 19:11 -!- saggi (purple): has joined #vdsm 19:12 -!- amuller (Assaf Muller): has joined #vdsm 19:23 < osvoboda> amuller, celebdor, danken: Bonding options reiterated :-) 19:27 <@danken> osvoboda: thanks for the notice! 19:28 <@danken> osvoboda: please rebase your patch 19:28 <@danken> vdsm/sos/vdsm.py.in vdsm/storage vdsm/vdsm vdsm_api vdsm_hooks vdsm_reg lib/vdsm/netlink.py:158:9: E123 closing bracket does not match indentation of opening bracket's line 19:28 <@danken> toni already hacked that. 19:28 < osvoboda> danken: Okay. 19:28 <@danken> osvoboda: not urgent 19:29 <@danken> after the next review 19:40 -!- ykaplan: has quit [Ping timeout: 272 seconds] 19:46 -!- amuller: has quit [Quit: Konversation terminated!] 19:52 <@danken> osvoboda: reviewed 19:52 < osvoboda> Thanks, already on it :-) 19:58 < osvoboda> danken: As can be seen in netinfoTests.py, netinfo.bondOpts() actually returns a dictionary of lists {'miimon'::}, I believe I have to ''.join() them for the legacy ['cfg']['BONDING_OPTS']. 20:06 -!- xaviern (Xavier): has joined #vdsm 20:53 -!- apahim: has quit [Ping timeout: 240 seconds] 20:57 -!- #vdsm,#theforeman xaviern: has quit [Ping timeout: 265 seconds] 20:57 -!- nsoffer (Nir Soffer): has joined #vdsm 21:02 -!- apahim (Amador Pahim): has joined #vdsm 21:09 -!- dougsland (Douglas): has joined #vdsm 21:30 -!- AlbertoSilva (AlbertoSilva): has joined #vdsm 21:30 -!- AlbertoSilva (AlbertoSilva): has joined #vdsm 21:39 -!- nsoffer: has quit [Read error: Connection reset by peer] 21:50 -!- xaviern (Xavier): has joined #vdsm 21:56 -!- #vdsm,#theforeman xaviern: has quit [Ping timeout: 252 seconds] 22:02 -!- osvoboda: has quit [Ping timeout: 265 seconds] 22:29 -!- fabiand: has quit [Quit: Verlassend] 22:35 -!- moolit: has quit [Quit: Leaving.] 23:29 -!- moolit (purple): has joined #vdsm 23:47 -!- moolit1 (purple): has joined #vdsm 23:47 -!- moolit: has quit [Read error: Connection reset by peer] 23:57 -!- osvoboda (Ondřej Svoboda): has joined #vdsm --- Log closed Wed Feb 19 00:00:21 2014