| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-11-08 | |||
| 14:23:39 | sean-k-mooney | why the peers group | |
| 14:23:45 | artom | Which do not get picked up on the compute host: https://zuul.opendev.org/t/openstack/build/eeae2903b14c456793766e79efd6916d/log/compute/logs/devstacklog.txt.gz#9048 | |
| 14:24:44 | sean-k-mooney | the whitebox plugin shoudl only need to be installed on teh contoler | |
| 14:25:14 | artom | We need it at least part of the setup to run on the compute host as well, for the SSH keys | |
| 14:25:46 | artom | 'cuz by default tempest runs as the 'tempest' user and it can't SSH anywhere | |
| 14:26:24 | sean-k-mooney | ok but you could do the ssh key setup as an ansible task | |
| 14:27:02 | sean-k-mooney | anyway so just trying to understand what proably you are hitting | |
| 14:27:16 | sean-k-mooney | you are not seeing the values take effect | |
| 14:27:19 | artom | I guess I could eh? | |
| 14:27:25 | artom | No, they're not being set | |
| 14:27:36 | sean-k-mooney | normaly we use subnode for devstack stuf | |
| 14:27:51 | sean-k-mooney | peers is only used to set up a vxlan tunnel between teh nodes | |
| 14:28:09 | artom | Should it matter? As long as the 'compute' node is in the group, which it is... | |
| 14:28:23 | artom | I'm using your own node definitions from your multi nfv job :) | |
| 14:28:33 | sean-k-mooney | it depends on how the ansible roles is written | |
| 14:28:50 | sean-k-mooney | yes and i alsye used subnode not peer | |
| 14:29:02 | sean-k-mooney | *always | |
| 14:30:15 | artom | One argument against using Ansible is that paramiko is finicky as fsck for its keys | |
| 14:30:15 | sean-k-mooney | by the way we only want the nodeset to be defiend in onepalce | |
| 14:30:29 | sean-k-mooney | im not sure about that | |
| 14:30:48 | sean-k-mooney | i have never had issues using keys creates with ssh-keygen | |
| 14:31:00 | artom | Well, maybe newer versions | |
| 14:31:04 | artom | But we're using 2.6.0 in the gate | |
| 14:31:20 | sean-k-mooney | older version fo ansibel are fine too | |
| 14:31:47 | artom | I'm kinda scared of going back | |
| 14:32:10 | artom | See, the exception that it gives has no bearing on reality, and I just discovered that recently | |
| 14:32:33 | artom | Failure to auth makes it throw a "not a valid OpenSSH key" exception | |
| 14:32:42 | artom | So maybe you're right, and I was lead astray by that | |
| 14:32:57 | artom | The real problem being authrozied_keys wasn't in place properly or something | |
| 14:33:15 | sean-k-mooney | ya likely | |
| 14:33:31 | sean-k-mooney | by the way are you are that you can see the renderd inventory in the job output | |
| 14:33:33 | sean-k-mooney | https://zuul.opendev.org/t/openstack/build/eeae2903b14c456793766e79efd6916d/log/zuul-info/inventory.yaml | |
| 14:34:19 | sean-k-mooney | artom: i think the problem is the subnodes devstack_localrc | |
| 14:34:33 | sean-k-mooney | is being used instead of peers | |
| 14:34:53 | sean-k-mooney | which is why i used subnode in my jobs | |
| 14:35:28 | sean-k-mooney | if you use subnode it will merge the section form the parent and the ones in your job | |
| 14:35:57 | sean-k-mooney | with your definition taking precidence | |
| 14:37:37 | artom | sean-k-mooney, oh, thanks! | |
| 14:37:50 | artom | Ugh, I know in my heart that ansible is the right way to do it | |
| 14:38:07 | artom | But I really don't want to revisit that | |
| 14:38:43 | mriedem | sean-k-mooney: you might have thoughts on helping the person in the mailing list trying to live migrate a vm between different ml2 backends on rocky which should be using the port binding stuff | |
| 14:38:47 | sean-k-mooney | just move the bits and you can leave devstack do it | |
| 14:38:53 | mriedem | i realize we don't have any troubleshooting docs for that | |
| 14:39:19 | sean-k-mooney | mriedem: i had not seen that but ill take a look. | |
| 14:39:28 | sean-k-mooney | it does not really work that well | |
| 14:39:37 | sean-k-mooney | there are still some gaps i think | |
| 14:40:07 | sean-k-mooney | it shoudl work betwwen thigns like ml2/odl and ml2/ovs | |
| 14:40:07 | mriedem | i'm pretty sure you identified several gaps at the denver ptg after rocky which are still open | |
| 14:40:21 | sean-k-mooney | ya i think we adressed 1 or two | |
| 14:40:23 | mriedem | e.. https://review.opendev.org/#/c/594527/ | |
| 14:41:04 | sean-k-mooney | ya that was how we found that | |
| 14:41:38 | sean-k-mooney | if your going between linux bridge and ovs there as also an issue where we were not passing hte bridge name | |
| 14:41:53 | sean-k-mooney | and a few other minor things | |
| 14:46:23 | dansmith | lyarwood: have we not moved to using the qemu-integrated iscsi driver yet? I was thinking we had | |
| 14:47:01 | lyarwood | dansmith: nope, I actually nuked the code for that a while ago as nothing was using it. | |
| 14:47:11 | dansmith | huh, okay | |
| 14:47:35 | lyarwood | dansmith: the original implentation was pretty bad, I've got a note to look at reintroducing it if there was a need somewhere. | |
| 14:48:06 | lyarwood | https://review.opendev.org/#/c/668750/ | |
| 14:48:10 | dansmith | I thought we were supposed to be moving towards it because it's vastly better and simpler than attaching those to the host | |
| 14:48:10 | lyarwood | ^ was the removal btw | |
| 14:49:01 | dansmith | ah recent | |
| 14:49:28 | lyarwood | it has been a long 4 months downstream ^_^ | |
| 14:49:44 | lyarwood | dansmith: what prompted your question btw? | |
| 14:50:03 | dansmith | lyarwood: this volume cache spec that I don't like | |
| 14:50:23 | mriedem | i got tim bell's attention https://review.opendev.org/#/c/566473/ | |
| 14:50:25 | lyarwood | ah | |
| 14:54:25 | sean-k-mooney | mriedem: it would be nice to see teh pending state become a thing and the reaper service either become a nova deliverable in its own repo or be broght in tree | |
| 14:54:45 | mriedem | double wrong | |
| 14:54:56 | sean-k-mooney | no? | |
| 14:54:56 | mriedem | first, you have to do what i said in there and land the rebuilding from cell0 piece | |
| 14:55:03 | mriedem | ignore pending for now | |
| 14:55:16 | mriedem | second, there was agreement that the reaper thing would be external to nova so we didn't need to own it | |
| 14:55:20 | mriedem | that's what the notifications are for | |
| 14:55:26 | sean-k-mooney | oh i taught that was already landed | |
| 14:55:54 | sean-k-mooney | oh ok | |
| 14:56:00 | mriedem | 1. build fails, goes to cell0, send notification. 2. external service consumes notification, purges spot instances based on it's own policy that nova doesn't care about, 3. external service triggers rebuild from cell0 on non-spot instance | |
| 14:56:11 | sean-k-mooney | well it would nice to be able to do premtiable isntance with openstack | |
| 14:56:21 | mriedem | sure, i agree :) | |
| 14:56:28 | sean-k-mooney | i guess i dont mind if its in or out of nova ownership | |
| 14:56:31 | mriedem | otherwise i wouldn't have spent time reviewing it | |
| 14:57:26 | mriedem | but it's also complicated enough that unless a core takes it over, it can't just be a thing thrown over the wall that gets worked on 1 day a week | |
| 14:58:01 | sean-k-mooney | ya, internally it would proably still land on the compute team to manage at redhat | |
| 14:58:03 | mriedem | with all the SIG hoopla and people that want stuff, there is always a laughable amount of effort put into a more than half day per week developer to push things | |
| 14:58:39 | mriedem | i'm not blaming theo, he did what he could and i'm sure he had other priorities downstream, especially since they already had that running in production | |
| 14:59:40 | sean-k-mooney | out of interest this was started around or before placement was a thing right | |
| 15:00:06 | sean-k-mooney | do the premtiable instace actully get modeled in placemnet? | |
| 15:00:28 | sean-k-mooney | im just wondering how it woudl interact with unified limits | |
| 15:02:42 | sean-k-mooney | they proably do have allocation as normal but i dont think they current unified lmits proposal could partion the quotas based on normal instance or premtiable at least at present | |
| 15:03:02 | sean-k-mooney | i think the workaound for that was to have a seperate project for each | |
| 15:04:19 | mriedem | don't know, been talking about spot instances and placement forever | |
| 15:04:28 | mriedem | and spot instances has changed many times over the years | |
| 15:04:41 | mriedem | doesn't really matter to us (nova) because the reaper is external and that's why we want it external | |
| 15:04:52 | mriedem | it's on the external service to define what the policy is | |
| 15:05:27 | mriedem | btw, if red hat were to take this over, at least the rebuild from cell0 part, i think melwitt would be a good person to take that over, since she's at least familiar with the spec | |
| 15:06:03 | mriedem | a lot of that piece is really just stuff that overlaps with the cross-cell work, which is moving data between cells and a buttload of testing | |
| 15:06:11 | melwitt | efried: fyi gibi_ptg's +2 on func test sleep removal now https://review.opendev.org/692252 | |
| 15:06:31 | efried | melwitt: ack, should get to that before lunch. | |
| 15:07:28 | mriedem | melwitt: can you rev the bug on that to be bug 1783565? | |
| 15:07:29 | openstack | bug 1783565 in OpenStack Compute (nova) "ServerGroupTestV21.test_evacuate_with_anti_affinity_no_valid_host intermittently fails with "Instance compute service state on host2 expected to be down, but it was up."" [Undecided,Invalid] https://launchpad.net/bugs/1783565 | |
| 15:07:30 | mriedem | your new one is a dupe | |