openstackgerrit | OpenStack Proposal Bot proposed openstack/dragonflow master: Updated from global requirements https://review.openstack.org/497025 | 05:48 |
---|---|---|
oanson | Morning | 06:11 |
dimak | Good morning | 06:23 |
dimak | Gate's still bad? | 06:23 |
openstackgerrit | Yuval Brik proposed openstack/dragonflow master: Redis Driver Rewrite [WIP] https://review.openstack.org/496299 | 06:28 |
oanson | Checking now. Did a recheck on a docs patch | 06:30 |
yuval | good morning | 07:01 |
oanson | o/ | 07:01 |
*** yamamoto has joined #openstack-dragonflow | 08:10 | |
oanson | dimak, looks like still broken | 08:16 |
dimak | oanson, yes.. :( | 08:16 |
oanson | Stacking locally to see if I can reproduce | 08:16 |
lihi | It seems that the gate is broken also without test_reconnect_of_controller | 08:40 |
*** saggi- has joined #openstack-dragonflow | 08:45 | |
*** saggi has quit IRC | 08:46 | |
*** saggi- is now known as saggi | 08:46 | |
*** Natanbro has joined #openstack-dragonflow | 08:50 | |
yuval | I broke nothing, I swear | 08:58 |
yuval | ¯\_(ツ)_/¯ | 08:58 |
openstackgerrit | Yuval Brik proposed openstack/dragonflow master: Redis Driver Rewrite [WIP] https://review.openstack.org/496299 | 09:07 |
*** yamamoto has quit IRC | 09:16 | |
*** yamamoto has joined #openstack-dragonflow | 09:17 | |
*** yamamoto has quit IRC | 09:27 | |
*** yamamoto has joined #openstack-dragonflow | 09:29 | |
*** yamamoto has quit IRC | 09:30 | |
evrardjp | hey | 09:48 |
evrardjp | oanson: could you ping us when you branch pike? | 09:48 |
oanson | Sure. | 09:49 |
oanson | Planned to happen on Friday, 1st of September :) | 09:49 |
evrardjp | ok thanks for the info! | 09:49 |
evrardjp | (so you know: we'll have to change the version in our openstack-ansible-os_neutron / defaults/main.yml -> dragonflow_git_install_branch to stable/pike) | 09:50 |
*** yamamoto has joined #openstack-dragonflow | 10:31 | |
*** yamamoto has quit IRC | 10:36 | |
*** Natanbro has quit IRC | 11:10 | |
openstackgerrit | Yuval Brik proposed openstack/dragonflow master: Redis Driver Rewrite [WIP] https://review.openstack.org/496299 | 11:25 |
*** yamamoto has joined #openstack-dragonflow | 11:33 | |
*** yamamoto has quit IRC | 11:38 | |
oanson | What I have so far: | 11:43 |
oanson | Reproduces on my Xenial VM. I can publish the local.conf if needed | 11:43 |
oanson | Looks like the ovsdb timeout causes future calls to vswitch_impl to hang | 11:43 |
oanson | Thus locking the main thread | 11:43 |
oanson | dimak lihi irenab ^^^ | 11:44 |
dimak | oanson, lihi and I are looking at ovsdbapp too | 11:45 |
dimak | Anything specific that was added between 040 and 050? | 11:46 |
oanson | Definitely worth while. Anything OVSDB related [should] go through there | 11:46 |
oanson | Let me check | 11:46 |
dimak | ovs itself if fully responsive to I don't think its that | 11:47 |
oanson | That's what I see too. When did we move over to 0.5.0 ? | 11:48 |
dimak | yesterday | 11:48 |
dimak | actually the day before in the evening | 11:49 |
dimak | upper constraints file in the requirements repo was updated | 11:49 |
oanson | Right | 11:50 |
irenab | oanson: https://review.openstack.org/496742 | 12:07 |
oanson | irenab, I don't think that's related. I don't see where we try to get this attribute, and the exception says there's a timeout. | 12:09 |
*** kkxue has joined #openstack-dragonflow | 12:14 | |
oanson | In any case, we use the attribute ovsdb_connection to store the relevant data (impl_idl.py, line 124) | 12:17 |
dimak | oanson, did you try cherrypicking the commit and rerunning? | 12:24 |
oanson | No | 12:24 |
dimak | If not, we can try tha | 12:24 |
dimak | t | 12:24 |
oanson | But I have a different suspect | 12:24 |
oanson | You do that. I'll revert a suspicious commit and try something too | 12:24 |
*** kkxue has quit IRC | 12:24 | |
dimak | Related to the timing out DB find command? | 12:25 |
oanson | Yes. | 12:27 |
oanson | There was a patch to support windows blocking which changes things around a bit | 12:27 |
oanson | But reverting it doesn't help :( | 12:28 |
*** yamamoto has joined #openstack-dragonflow | 12:34 | |
oanson | I think this commit is the culprit: 903dd1df26c7d8babc7cec7dc920c2bb850bd841 Add OVN_Southbound API support | 12:35 |
oanson | lihi, dimak ^^^ | 12:37 |
dimak | Confirmed? | 12:37 |
oanson | In ovsdbapp | 12:37 |
dimak | Looking now | 12:37 |
oanson | I started from 0.5.0 and stepped back. Stepping back from this commit made the test pass (it failed consistently before) | 12:37 |
oanson | So, empirically | 12:37 |
dimak | good old scientific method | 12:38 |
*** yamamoto has quit IRC | 12:41 | |
oanson | Yeah. Also, funny thing, if you don't start a connection, it doesn't work... :( https://review.openstack.org/#/c/481779/12/ovsdbapp/schema/open_vswitch/impl_idl.py@78 | 12:41 |
oanson | Uploading a potential fix in a mo | 12:42 |
dimak | I see that now | 12:43 |
openstackgerrit | Omer Anson proposed openstack/dragonflow master: gate-fix: Start ovsdb connection https://review.openstack.org/497230 | 12:43 |
oanson | The most annoying thing is, I didn't review this one since I said 'Nah! It's OVN related stuff. What would I know?' | 12:44 |
dimak | Tested it locally, stuff seems better | 12:45 |
oanson | The fix, or the revert? | 12:46 |
lihi | the fix | 12:46 |
oanson | Cool! | 12:46 |
lihi | oanson, this is what's happen when you neglect your duties | 12:47 |
oanson | Yep. :( | 12:47 |
* oanson bows his head in shame | 12:48 | |
*** yamamoto has joined #openstack-dragonflow | 13:36 | |
*** yamamoto has quit IRC | 13:42 | |
openstackgerrit | Yuval Brik proposed openstack/dragonflow master: Redis Driver Rewrite [WIP] https://review.openstack.org/496299 | 13:44 |
*** mlavalle has joined #openstack-dragonflow | 14:00 | |
openstackgerrit | Omer Anson proposed openstack/dragonflow master: Add nb db migration mechanism and apply to lswitch unique_key https://review.openstack.org/401210 | 14:26 |
*** yamamoto has joined #openstack-dragonflow | 14:38 | |
*** yamamoto has quit IRC | 14:44 | |
oanson | dimak, lihi, the gate fix patch passed gate. But I have a concession - I think setting self.ovsdb_connection is wrong. Do we want to merge and I'll fix it in a different patch, or are we patient for another 2 hours? | 14:56 |
dimak | We're good | 15:12 |
dimak | Lets try to fix it correctly | 15:12 |
dimak | oanson, ^^^ | 15:13 |
oanson | Sure. Will upload new patch momentarily | 15:14 |
openstackgerrit | Omer Anson proposed openstack/dragonflow master: gate-fix: Do not set ovsdb_connection as attribute https://review.openstack.org/497230 | 15:15 |
oanson | dimak, lihi, keep your fingers crossed ^^^^ | 15:15 |
*** yamamoto has joined #openstack-dragonflow | 15:40 | |
*** yamamoto has quit IRC | 15:45 | |
*** openstackgerrit has quit IRC | 16:04 | |
*** yamamoto has joined #openstack-dragonflow | 16:41 | |
*** yamamoto has quit IRC | 16:47 | |
*** yamamoto has joined #openstack-dragonflow | 17:43 | |
*** yamamoto has quit IRC | 17:48 | |
*** yamamoto has joined #openstack-dragonflow | 18:44 | |
*** yamamoto has quit IRC | 18:50 | |
*** yamamoto has joined #openstack-dragonflow | 19:46 | |
*** yamamoto has quit IRC | 19:51 | |
*** yamamoto has joined #openstack-dragonflow | 19:51 | |
*** openstackgerrit has joined #openstack-dragonflow | 20:23 | |
openstackgerrit | Omer Anson proposed openstack/dragonflow master: gate-fix: Do not set ovsdb_connection as attribute https://review.openstack.org/497230 | 20:23 |
*** yamamoto has quit IRC | 20:45 | |
*** yamamoto has joined #openstack-dragonflow | 21:46 | |
*** yamamoto has quit IRC | 21:51 | |
*** yamamoto has joined #openstack-dragonflow | 22:47 | |
*** yamamoto has quit IRC | 22:53 | |
*** mlavalle has quit IRC | 23:43 | |
*** yamamoto has joined #openstack-dragonflow | 23:49 | |
*** yamamoto has quit IRC | 23:55 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!