| rm_work | noticing another issue with rocky | 01:57 |
|---|---|---|
| rm_work | if you try to run it on py36/py37 it explodes due to eventlet issues | 01:57 |
| rm_work | rocky pins eventlet==0.20.0 | 01:58 |
| rm_work | but there's a bug in that version which causes a recusion-depth explosion | 01:58 |
| rm_work | see https://bugs.launchpad.net/manila/+bug/1788253 | 01:58 |
| openstack | Launchpad bug 1788253 in Manila "SSLContext infinite recursion in Python 3.6 " [High,Fix released] - Assigned to Goutham Pacha Ravi (gouthamr) | 01:58 |
| rm_work | need eventlet==0.21.0 | 01:59 |
| rm_work | but apparently just having that installed breaks some unit tests O_o | 01:59 |
| rm_work | which seem very unrelated... | 01:59 |
| rm_work | not a FULL log (there's more) but for example: http://paste.openstack.org/show/790654/ | 01:59 |
| rm_work | if i have eventlet 0.20.0 these don't fail :/ | 02:00 |
| rm_work | eandersson: i have an idea to fix the eventlet tests: drop eventlet :D | 02:07 |
| rm_work | why are you still using eventlet in 2020 T_T | 02:08 |
| rm_work | anyway, i can replicate this locally, but i don't understand why these tests only fail with a different eventlet version | 02:23 |
| rm_work | are they being SKIPPED or something? they should NEVER work on py3 | 02:23 |
| rm_work | regardless of eventlet | 02:23 |
| rm_work | with eventlet 0.20.0: Ran 391 tests in 7.614s (-3.517s) | 02:25 |
| rm_work | with eventlet 0.25.1: Ran 391 tests in 8.976s (+1.528s) | 02:26 |
| rm_work | so same | 02:26 |
| rm_work | how the heck do these pass in py3 | 02:26 |
| rm_work | wtf | 03:04 |
| rm_work | https://github.com/openstack/designate/blob/stable/rocky/designate/mdns/handler.py#L346 | 03:06 |
| rm_work | py3.7, eventlet 0.20.0, `type(q_rrset.name.to_text()) -> bytes` | 03:06 |
| rm_work | py3.7, eventlet 0.25.1, `type(q_rrset.name.to_text()) -> str` | 03:06 |
| rm_work | HOW DOES THE EVENTLET VERSION AFFECT THIS | 03:06 |
| rm_work | i hate EVERYTHING | 03:07 |
| openstackgerrit | Adam Harwell proposed openstack/designate stable/rocky: Eventlet bump + properly decode strings in py3x https://review.opendev.org/712830 | 03:46 |
| *** bnemec has quit IRC | 05:19 | |
| *** bnemec has joined #openstack-dns | 05:28 | |
| mugsie | well, that was a trip to watch :) | 10:26 |
| mugsie | rm_work: please, if you have time I would love to do the rm -rf eventlet | 10:27 |
| rm_work | :D | 10:42 |
| rm_work | oh hey: | 10:42 |
| rm_work | "Managed Resource Tenant ID is not properly configured" | 10:42 |
| rm_work | coming fron central... is that a dealbreaker? or just a warning? | 10:42 |
| rm_work | guessing it's ignorable | 10:45 |
| mugsie | it is ignorable, but shoud really be set to a service project id | 10:49 |
| rm_work | hmm this is weird | 10:54 |
| rm_work | `designate-manage powerdns sync` is hanging | 10:55 |
| rm_work | like it's not getting a reply back from central | 10:55 |
| mugsie | rm_work: which version of pdns are you using? | 11:02 |
| rm_work | hmmm | 11:02 |
| mugsie | is the driver pdns or pdns4 ? | 11:02 |
| rm_work | hmmmmmmmm | 11:02 |
| rm_work | (stalling while i get my yubi-key auth done) | 11:03 |
| mugsie | :) | 11:03 |
| rm_work | err i don't see that in the config | 11:03 |
| rm_work | i feel like that should be set, but it isn't. lol | 11:05 |
| mugsie | it should be in pools.yaml | 11:05 |
| rm_work | oh | 11:05 |
| rm_work | i know nothing about designate :D | 11:05 |
| * mugsie cries | 11:05 | |
| mugsie | and i spent all that time learning about octavia | 11:06 |
| rm_work | :S | 11:06 |
| rm_work | `type: powerdns`? | 11:06 |
| rm_work | though this may just be a rabbit issue in the end, trying to prove out whether that's the case or not | 11:07 |
| mugsie | yeah, then it is probably a rmq issue | 11:07 |
| rm_work | yeah looks like central never picks up any messages :( | 11:10 |
| rm_work | err, for this | 11:10 |
| rm_work | apparently other messages are fine :D | 11:10 |
| mugsie | is your designate config ina weird place? | 11:13 |
| mugsie | (i.e. not /etc/designate/designate.conf ?) | 11:13 |
| rm_work | nope | 11:18 |
| rm_work | tried passing it explicitly | 11:18 |
| rm_work | super weird, just hanging on sending the message to rmq it seems... other stuff is sending messages fine <_< | 11:19 |
| *** nicolasbock has joined #openstack-dns | 11:25 | |
| *** stingrayza has quit IRC | 12:01 | |
| *** stingrayza has joined #openstack-dns | 12:05 | |
| *** nicolasbock has quit IRC | 12:07 | |
| *** nicolasbock has joined #openstack-dns | 12:07 | |
| rm_work | yeah seems like something with eventlet >_< | 12:35 |
| rm_work | hanging on sleep() | 12:36 |
| rm_work | maybe... | 12:36 |
| *** bnemec is now known as beekneemech | 14:10 | |
| *** openstackgerrit has quit IRC | 16:31 | |
| *** openstackgerrit has joined #openstack-dns | 16:41 | |
| *** ChanServ sets mode: +v openstackgerrit | 16:41 | |
| openstackgerrit | Sean McGinnis proposed openstack/python-designateclient master: Use unittest.mock instead of third party mock https://review.opendev.org/713021 | 16:41 |
| openstackgerrit | Sean McGinnis proposed openstack/python-designateclient master: Drop py27 support https://review.opendev.org/713044 | 18:29 |
| openstackgerrit | Sean McGinnis proposed openstack/python-designateclient master: Use unittest.mock instead of third party mock https://review.opendev.org/713021 | 18:30 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!