| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-01-16 | |||
| 18:13:01 | ildikov | mriedem: thanks!!! | |
| 18:14:08 | openstackgerrit | Matt Riedemann proposed openstack/nova: WIP Support cinder v3 API https://review.openstack.org/420201 | |
| 18:20:33 | openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Reconnect volumes and encryptors during a hard reboot https://review.openstack.org/400384 | |
| 18:24:53 | mriedem | dansmith: so i'm kicking around ideas of how to make the multinode job run multiple cells, | |
| 18:25:12 | mriedem | i think at least part of that is calling map_cell_and_hosts from the subnode | |
| 18:25:22 | dansmith | mriedem: multiple cellsv2 cells you mean? | |
| 18:25:27 | mriedem | yeah | |
| 18:25:33 | mriedem | so host 1 is control + n-cpu, | |
| 18:25:38 | mriedem | host 2 (subnode) is just n-cpu | |
| 18:25:47 | mriedem | but both hosts are going to be a cell | |
| 18:26:41 | mriedem | or a combo of create_cell and discover_hosts | |
| 18:26:44 | mriedem | on the subnode | |
| 18:27:30 | dansmith | doesn't have to be on the subnode specifically, | |
| 18:27:38 | dansmith | but yeah, create_cell will make it more explicit I think | |
| 18:29:14 | mriedem | so currently we just create the dbs on the api node https://github.com/openstack-dev/devstack/blob/master/lib/nova#L683 | |
| 18:29:18 | mriedem | which is fine | |
| 18:29:52 | melwitt | I was thinkin once we have create_cell, we'd remove map_cell_and_hosts since that's really just create_cell + discover_hosts. to make things not overlap and be simpler | |
| 18:30:04 | dansmith | yep | |
| 18:30:12 | mriedem | then we do the mappings later in https://github.com/openstack-dev/devstack/blob/master/lib/nova#L948 | |
| 18:30:53 | mriedem | which is called later here https://github.com/openstack-dev/devstack/blob/master/stack.sh#L1368 | |
| 18:30:59 | mriedem | so we only do that stuff if on the api node | |
| 18:31:18 | mriedem | so i think i can put an else in there for when n-api isn't running (subnode) | |
| 18:31:21 | mriedem | and n-cpu is enabled, | |
| 18:31:49 | mriedem | but would probably want the subnode cell db on the api node, i'm not sure if it matters for CI? | |
| 18:31:58 | mriedem | given we're hitting OOM errors in the gate right now, | |
| 18:32:03 | mriedem | the cell db on the subnode is probably better | |
| 18:32:11 | dansmith | doesn't need to actually be a different place | |
| 18:32:21 | mriedem | yeah i know, | |
| 18:32:30 | mriedem | but devstack isn't topology aware, | |
| 18:32:40 | mriedem | and i'm not entirely sure how all of the subnode stuff gets setup by devstack-gate | |
| 18:32:51 | mriedem | so i'm not sure that when setting up the api node, we know about the subnode yet | |
| 18:33:05 | mriedem | but i guess maybe that doesn't matter? | |
| 18:33:42 | mriedem | we just have a db somewhere, and when we create the cell for the subnode we point that cell at that db connection | |
| 18:34:46 | dansmith | well, | |
| 18:35:01 | dansmith | the thing I ran into earlier trying to do something else was that the subnode didn't have all the stuff it needed to do certain things | |
| 18:35:05 | dansmith | like, not all the config | |
| 18:35:22 | dansmith | don't recall what exactly, but I couldn't start n-api on the subnode | |
| 18:35:29 | dansmith | might've been keystone and catalog related actually | |
| 18:36:42 | mriedem | this is what we have on the subnode http://logs.openstack.org/75/396775/39/check/gate-tempest-dsvm-neutron-multinode-full-ubuntu-xenial-nv/f82fd9f/logs/subnode-2/etc/nova/nova.conf.txt.gz | |
| 18:40:04 | mriedem | somehow the subnode is getting the variables for the db host | |
| 18:41:14 | mriedem | ah d- | |
| 18:41:18 | mriedem | *d-g my old friend | |
| 18:44:36 | mriedem | alright time to hacky hacky | |
| 18:50:06 | jaosorior | gibi: you around? | |
| 18:50:35 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/420844 | |
| 18:59:57 | openstackgerrit | Jay Pipes proposed openstack/nova: placement: RT now adds proper Ironic inventory https://review.openstack.org/404472 | |
| 18:59:59 | openstackgerrit | Jay Pipes proposed openstack/nova: rt: use a single ResourceTracker object instance https://review.openstack.org/398473 | |
| 19:00:01 | jaypipes | mriedem: ^^ | |
| 19:00:21 | mriedem | que bueno | |
| 19:01:26 | cdent | good timing jaypipes my to do list just rolled around to those | |
| 19:01:34 | jaypipes | coolio. | |
| 19:01:41 | mriedem | jaypipes: were you going to also update _setup_pci_tracker to pass the compute_node object in ther? | |
| 19:01:43 | mriedem | *there | |
| 19:01:54 | jaypipes | mriedem: ah, shitbuckets. | |
| 19:01:56 | jaypipes | one sec. | |
| 19:02:34 | mriedem | otherwise it looks super hot | |
| 19:02:35 | mriedem | like, omg | |
| 19:08:03 | mriedem | melwitt: good point on https://review.openstack.org/#/c/332713/14/nova/cmd/manage.py@1432 | |
| 19:08:10 | mriedem | i obviously didn't think about that | |
| 19:09:09 | mriedem | heh i even said that in PS9: | |
| 19:09:10 | mriedem | "I think I'd like to see us first check to see if a cell mapping exists with the given name (or combination of name/transport_url/database_connection) and if we already have a match, we just exit without doing anything (not a failure)." | |
| 19:10:17 | melwitt | mriedem: yeah, I've been admittedly confused about whether we return 0 for a no-op or not among all these commands. it seems like it hasn't been consistent | |
| 19:11:04 | melwitt | like simple_cell_setup returns non-zero if it didn't create a cell mapping, but map_cell_and_hosts returns 0 in the same case I think | |
| 19:11:25 | mriedem | yeah bauzas has a patch up to fix that actually | |
| 19:11:53 | mriedem | another thing with this create_cell command is you can't create 2 cells with different names but the same db/mq urls | |
| 19:12:02 | mriedem | i'm not sure if we care about that right now, and it could be built in later if we did | |
| 19:12:26 | mriedem | seems we should have had a unique constraint on the name/db/transport combo in the schema | |
| 19:12:40 | melwitt | yeah, I thought about that and figured we shouldn't probably allow that case anyway (different name, same urls) | |
| 19:13:54 | dansmith | we should allow duplicate urls of each type at the very least | |
| 19:14:14 | dansmith | not sure it makes sense to allow complete dupes though, other than for testing | |
| 19:15:30 | mriedem | dansmith: yeah the check in create_cell right now is db AND mq urls matching an existing cell | |
| 19:15:32 | mriedem | that's a failure | |
| 19:15:52 | mriedem | which i had earlier said i thought should just be a noop (don't create a new cell), but guess i forgot that i asked for that before +2ing | |
| 19:16:02 | openstackgerrit | Jim Rollenhagen proposed openstack/nova: Move to tooz hash ring implementation https://review.openstack.org/399022 | |
| 19:16:11 | jaypipes | mriedem: just re-running tests locally. should be pushed within 15minutes. | |
| 19:16:39 | mriedem | ok push it real good please | |
| 19:17:43 | dansmith | mriedem: well, as melwitt said, returning zero for "I didn't do anything, but it's there like you asked" will be good for some people and bad for others that think they can test for existence by trying to re-create and check for error | |
| 19:18:03 | dansmith | mriedem: so my opinion is that we not make a "create this thing" operation idempotent and instead do what you asked | |
| 19:18:12 | mriedem | right, and since i'm frankly tired of trying to predict how all of this is going to be used, i'm going to give up | |
| 19:20:26 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/420844 | |
| 19:21:26 | mriedem | dansmith: don't forget about this guy https://review.openstack.org/#/c/420440/ | |
| 19:21:46 | dansmith | mriedem: jenkins says that is shit | |
| 19:22:10 | mriedem | it said that about the next in the series too which you +2ed | |
| 19:22:52 | dansmith | I hand-checked that one | |
| 19:23:23 | mriedem | i know | |
| 19:23:45 | mriedem | my detector has been honed over years of bullshit | |
| 19:33:56 | oomichi_ | mriedem: why not "nova-manage cell_v2 list" ? The other resources' commands are just "list" without resource name in sub commands | |
| 19:34:55 | mriedem | oomichi_: because we could have things like list_host_mappings or something, idk | |
| 19:35:07 | mriedem | oomichi_: and that's what we named it like for create_cell | |
| 19:35:58 | dansmith | I commented to that effect | |
| 19:36:09 | oomichi_ | mriedem: ah, I see "create_cell". I got it. | |
| 19:36:33 | oomichi_ | dansmith: Thanks, I see. it is nice to keep current one for consistent command line | |
| 19:37:49 | openstackgerrit | Jay Pipes proposed openstack/nova: placement: RT now adds proper Ironic inventory https://review.openstack.org/404472 | |
| 19:37:52 | openstackgerrit | Jay Pipes proposed openstack/nova: rt: use a single ResourceTracker object instance https://review.openstack.org/398473 | |
| 19:40:41 | melwitt | omg fixing unit tests is finding my bugs -_- | |
| 19:57:55 | melwitt | dansmith: are there any common mistakes/reasons why a lazy-load of flavor would try to occur even after specifying expected_attrs=['flavor']? | |
| 19:58:48 | dansmith | melwitt: hmm, not that I can think of | |
| 20:00:24 | slaweq | mriedem: hello | |
| 20:00:44 | openstack | Launchpad bug 1640502 in OpenStack Compute (nova) "Empty /etc/machine-id causes failure in unit tests" [Medium,In progress] - Assigned to Slawek Kaplonski (slaweq) | |
| 20:00:44 | slaweq | mriedem: I made small patch https://review.openstack.org/#/c/420176/ to fix https://bugs.launchpad.net/nova/+bug/1640502 | |