*** masber has joined #openstack-dns | 00:03 | |
*** masber has quit IRC | 00:03 | |
*** masber has joined #openstack-dns | 00:03 | |
*** masber has quit IRC | 00:04 | |
*** masber has joined #openstack-dns | 00:04 | |
*** masber has quit IRC | 00:04 | |
*** masber has joined #openstack-dns | 00:06 | |
*** masber has quit IRC | 00:06 | |
*** masber has joined #openstack-dns | 00:08 | |
*** trungnv has joined #openstack-dns | 01:33 | |
*** caowei has joined #openstack-dns | 02:23 | |
*** pcaruana has joined #openstack-dns | 06:31 | |
*** Tbone has joined #openstack-dns | 06:59 | |
Tbone | Morning all, I just started with a kolla deployment of designate and everything seemed to have gone well. I noticed however that the mdns service was listening on the public interface which is fine but the bind9 service was listening on the private network interface. bind9 should definitely listen on the public interface as well.. | 07:02 |
---|---|---|
Tbone | just wanted to make sure if anyone else deployed designate via kolla and experienced the same ? | 07:02 |
*** trungnv has quit IRC | 07:35 | |
*** trungnv has joined #openstack-dns | 07:35 | |
*** AlexeyAbashkin has joined #openstack-dns | 07:46 | |
*** ivve has quit IRC | 08:11 | |
*** salmankhan has joined #openstack-dns | 08:12 | |
*** MarkBaker has joined #openstack-dns | 08:23 | |
*** sapcc-bot4 has joined #openstack-dns | 09:00 | |
openstackgerrit | Nguyen Van Trung proposed openstack/designate master: Allow to check pep8 along with D000 https://review.openstack.org/567458 | 09:03 |
*** sapcc-bot3 has quit IRC | 09:04 | |
openstackgerrit | Nguyen Van Trung proposed openstack/designate master: Change copyright information for field.py https://review.openstack.org/565237 | 09:06 |
*** masber has quit IRC | 09:06 | |
*** salmankhan has quit IRC | 09:25 | |
*** salmankhan has joined #openstack-dns | 09:31 | |
openstackgerrit | Nguyen Van Trung proposed openstack/designate master: Allow to check pep8 along with D000 https://review.openstack.org/567458 | 09:40 |
*** abalutoiu has joined #openstack-dns | 09:47 | |
*** Tbone has quit IRC | 11:14 | |
*** livelace has joined #openstack-dns | 11:50 | |
*** Tbone has joined #openstack-dns | 11:54 | |
*** abalutoiu has quit IRC | 12:10 | |
*** abalutoiu has joined #openstack-dns | 12:23 | |
*** MarkBaker has quit IRC | 12:24 | |
*** abalutoiu_ has joined #openstack-dns | 12:36 | |
*** MarkBaker has joined #openstack-dns | 12:40 | |
*** abalutoiu has quit IRC | 12:40 | |
*** MarkBaker has quit IRC | 13:25 | |
*** MarkBaker has joined #openstack-dns | 13:37 | |
*** Tbone has quit IRC | 13:48 | |
*** beekneemech is now known as bnemec | 13:59 | |
*** MarkBaker has quit IRC | 14:21 | |
*** pcaruana has quit IRC | 15:02 | |
eandersson | Tbone there is a variable called dns_interface for designate-mdns to control which interface private vs. public designate should be using | 16:08 |
eandersson | As for bind take a look at this https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/designate/templates/named.conf.j2 | 16:09 |
eandersson | > #api_interface: "{{ network_interface }}" | 16:10 |
eandersson | It looks like you need to modify that variable to have bind listen to a public interface | 16:10 |
*** pcaruana has joined #openstack-dns | 16:20 | |
*** salmankhan has quit IRC | 16:31 | |
*** rfreire has joined #openstack-dns | 16:38 | |
*** AlexeyAbashkin has quit IRC | 16:57 | |
*** rfreire_ has joined #openstack-dns | 17:02 | |
*** rfreire is now known as Guest25989 | 17:02 | |
*** rfreire_ is now known as rfreire | 17:02 | |
*** mugsie has quit IRC | 17:06 | |
*** mugsie has joined #openstack-dns | 17:06 | |
*** mugsie has quit IRC | 17:08 | |
*** mugsie has joined #openstack-dns | 17:12 | |
*** mugsie has quit IRC | 17:12 | |
*** mugsie has joined #openstack-dns | 17:12 | |
*** sapd1 has joined #openstack-dns | 17:12 | |
*** rfreire has quit IRC | 17:27 | |
*** rfreire has joined #openstack-dns | 17:28 | |
eandersson | mugsie, for the Bad Action error | 17:49 |
eandersson | Do you think it's something that would be safe to ignore? as in catching and logging, as opposed to the current exception? | 17:49 |
mugsie | It can be an issue, I think | 17:50 |
mugsie | *think* | 17:50 |
mugsie | I think it might have something to do with OVO | 17:50 |
mugsie | but that is just my gut speaking | 17:50 |
eandersson | I see - I can try to dig a big deeper | 17:50 |
eandersson | For sure feels like a race condition | 17:51 |
mugsie | but - in theory, we should be able to recover this later anyway, so ignoring it may be fine | 17:51 |
eandersson | Yea - the lab looks fine, but of course that is a tiny deployment, not sure how this would look in a prod env | 17:52 |
mugsie | Yeah. I am more worried right now about the service status bug that is open | 17:53 |
eandersson | Ah yea been meaning to look into that one | 17:53 |
rfreire | <eandersson> mugsie, for the Bad Action error <- I've seen some in Netflix | 17:54 |
eandersson | Ah yea - the db thing is a thing | 17:55 |
eandersson | Just confirmed in my lab | 17:55 |
eandersson | count(*) 438576 | 17:55 |
mugsie | eandersson: https://media.giphy.com/media/h3LIsfXuPqALK/giphy.gif | 17:58 |
mugsie | rektide: Netflix? | 17:58 |
mugsie | s/rektide/rfreire/ | 17:58 |
rfreire | mugsie, bad action, bad actors.. | 17:58 |
rfreire | just a monday bad joke | 17:58 |
rfreire | <grins> | 17:58 |
mugsie | nice | 17:58 |
rfreire | mugsie, I deserved that giphy | 17:58 |
rfreire | that was awful bad | 17:59 |
*** ircuser-1 has joined #openstack-dns | 18:05 | |
*** idlemind has joined #openstack-dns | 18:27 | |
eandersson | mugsie, > [req-0f896b0f-a0d6-4edd-815a-15787ec31a07 None None] Could not find ServiceStatus: ServiceStatusNotFound: Could not find ServiceStatus | 18:28 |
mugsie | -_- | 18:29 |
eandersson | http://paste.openstack.org/show/720943/ | 18:29 |
eandersson | I'll throw in a quick PR | 18:31 |
*** salmankhan has joined #openstack-dns | 18:33 | |
mugsie | Yeah, this is a rocky issue | 18:40 |
mugsie | queens release for my stuff is fine | 18:40 |
eandersson | The above was from Queens | 18:40 |
eandersson | > designate==6.0.1 | 18:41 |
mugsie | uuuuu .... | 18:41 |
mugsie | graham@s003:/etc/designate$ dpkg -l designate | 18:41 |
mugsie | Desired=Unknown/Install/Remove/Purge/Hold | 18:41 |
mugsie | | Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend | 18:41 |
mugsie | |/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad) | 18:41 |
mugsie | ||/ Name Version Architecture Description | 18:41 |
mugsie | +++-===================================================-==============================-==============================-=========================================================================================================== | 18:41 |
mugsie | ii designate 1:6.0.0-0ubuntu1~cloud0 all OpenStack DNS as a Service - metapackage | 18:41 |
mugsie | oh, no way | 18:41 |
eandersson | It's only affecting worker/producer | 18:42 |
eandersson | Still trying to figure out exactly why | 18:42 |
eandersson | So this might be more than one issue | 18:47 |
eandersson | let me whipe the status database | 18:48 |
eandersson | lol can't delete the entries due to deadlock | 18:49 |
rfreire | \o/ | 18:50 |
eandersson | so I think what is happening is that at some point a duplicate entry gets created for the worker/producer | 18:52 |
eandersson | when that happens the sql query will never succeed, as the value is no longer unique | 18:52 |
eandersson | service_name + hostname should always be unique | 18:52 |
eandersson | so probably a schema issue | 18:53 |
eandersson | mugsie, you can probably reproduce this by creating a duplicate entry for one of the services | 18:54 |
mugsie | yeah, that seems legit | 18:59 |
*** sapd1 has quit IRC | 19:10 | |
*** rfreire has quit IRC | 19:25 | |
*** rfreire has joined #openstack-dns | 19:26 | |
*** rfreire has quit IRC | 19:34 | |
*** rfreire has joined #openstack-dns | 19:35 | |
*** rfreire has quit IRC | 19:47 | |
*** salmankhan has quit IRC | 19:48 | |
eandersson | mugsie, never touched the orm stuff before, you wanna do the patch? | 19:59 |
eandersson | We just need to do something like this I guess https://github.com/openstack/designate/blob/master/designate/storage/impl_sqlalchemy/migrate_repo/versions/082_unique_ns_record.py | 19:59 |
eandersson | This is unfortunately a two year old bug. | 20:01 |
eandersson | We should handle this in code as well. | 20:02 |
eandersson | maybe change find_service_status to find_service_statuses in central, and raise an exception if there are more than one | 20:03 |
*** pcaruana has quit IRC | 20:43 | |
openstackgerrit | Erik Olof Gunnar Andersson proposed openstack/designate master: [WIP] Add unqiue constraint to service_statuses https://review.openstack.org/568372 | 20:52 |
*** livelace has quit IRC | 20:54 | |
*** kiall has quit IRC | 20:54 | |
*** kiall has joined #openstack-dns | 20:55 | |
*** kiall has joined #openstack-dns | 20:55 | |
eandersson | mugsie, ^ | 21:03 |
*** bnemec has quit IRC | 21:17 | |
*** bnemec has joined #openstack-dns | 21:18 | |
openstackgerrit | Erik Olof Gunnar Andersson proposed openstack/designate master: [WIP] Add unqiue constraint to service_statuses https://review.openstack.org/568372 | 22:11 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!