fgimenez | hi Saviq, yesterday i took a look into this https://jenkins.qa.ubuntu.com/job/generic-deb-autopilot-runner-vivid-mako/2080/console during my vanguard turn | 11:15 |
---|---|---|
fgimenez | Saviq, i haven't confirmed it yet but it seems that the messages are coming from the dbus package http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/dbus-python/vivid/view/head:/dbus/proxies.py#L413 | 11:15 |
fgimenez | Saviq, this seems to be the loop http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/dbus-python/vivid/view/head:/dbus/proxies.py#L386 | 11:16 |
Saviq | fgimenez, oh, I'd think the exception would be raised in that case? | 11:17 |
fgimenez | Saviq, i'll try to confirm it this evening and ping you with the results | 11:17 |
Saviq | fgimenez, thanks | 11:17 |
=== MacSlow is now known as MacSlow|lunch | ||
=== MacSlow|lunch is now known as MacSlow | ||
fgimenez | elopio, it seems that Savic's issue https://jenkins.qa.ubuntu.com/job/generic-deb-autopilot-runner-vivid-mako/2080/console comes from the dbus library | 15:27 |
fgimenez | elopio, i've managed to reproduce something similar http://paste.ubuntu.com/10871535/ by modifying this function http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/dbus-python/vivid/view/head:/dbus/connection.py#L602 | 15:28 |
fgimenez | elopio, this happens when calling autopilot/introspection/_search.py#get_proxy_object_for_existing_process, but there's no exception raised, should ap be aware of any timeout? | 15:29 |
elopio | fgimenez: lots of information, that's good, thanks! | 15:32 |
elopio | and I don't know. I'm still parsing what you are saying. fgimenez: it would be better to send an email to veebers, with cc to the qa list. | 15:33 |
elopio | fgimenez: the issue is that unity is dying, right? What you are digging is to report a better error message in that case? or am I completely lost? | 15:34 |
fgimenez | elopio, yep, ap is trying to get the proxy object for the unity8 process, but it has already died, or is not accessible by dbus | 15:35 |
fgimenez | elopio, then it seems that dbus spends a lot of time asking the introspection tree | 15:36 |
fgimenez | elopio, locally and with a custom error_handler it doesn't take too much, in jenkins the job timeout is expired | 15:36 |
fgimenez | elopio, will put it together in an email, i'll try to understand first which objects are queried exactly in dbus after the first error | 15:41 |
elopio | fgimenez: nice. | 15:53 |
elopio | fgimenez: Saviq: about restarting as a workaround while you solve the unity hang, I think I prefer to keep the tests failing. | 15:58 |
Saviq | fgimenez, in our case unity8 is deadlocked in dbus (bug #1421009) | 15:59 |
ubot5 | bug 1421009 in qtbase-opensource-src (Ubuntu) "unity8 sometimes hangs on boot" [Critical,In progress] https://launchpad.net/bugs/1421009 | 15:59 |
Saviq | elopio, meh, but we might get ↑ fixed soon anyway | 16:00 |
elopio | Saviq: that would be perfect :) | 16:14 |
balloons | veebers, heya. So mhall is away in London this week, but apartly IS is running into some trouble with his request to land things (including the docs). Just wanted to give you a heads up | 20:29 |
veebers | balloons: ack thanks. Any idea what the issues are? | 20:32 |
balloons | veebers, my brain is failing me at this hour. Presumably they are noted in the ticket should you have a link handy | 20:35 |
balloons | we can inquire directly with IS; might be better | 20:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!