| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-07-18 | |||
| 13:53:12 | sean-k-mooney | mordred: well os-vif never used the install script because it broke my cygwin workflow at the time. | |
| 13:53:31 | tssurya | jaypipes: you appear to be the only spec core online now; whenever you get the time could you please have a look at https://review.openstack.org/#/c/581243/ ? It a very easy not fix follow-up on a merged spec | |
| 13:53:54 | tssurya | nit* | |
| 13:54:36 | mordred | sean-k-mooney: \o/ | |
| 13:57:11 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: Transform volume.usage notification https://review.openstack.org/580345 | |
| 13:59:23 | efried | JulienCornuwel: To be a bit more specific: your notification listener would conceivably start the cleanup process, and also register the device in some kind of cache ("reserved devices"). Add logic to your get_available_resource method to look at that cache and skip reporting any devices therein. | |
| 13:59:23 | efried | So when we make the move to placement-based device reporting, the latter logic would go into update_provider_tree instead of get_available_resource, but everything else would stay the same. | |
| 13:59:23 | efried | There may still be a race in this setup, because there may be a delay before get_available_resource/update_provider_tree is run. I suppose you could get around this by making your spawn method add the device to the "reserved" list, and then have your notification listener thingy only be responsible for removing it when done. | |
| 13:59:38 | openstackgerrit | sean mooney proposed openstack/os-vif master: add upper_constraitns support https://review.openstack.org/583582 | |
| 14:00:12 | sean-k-mooney | we should backport ^ | |
| 14:01:48 | sean-k-mooney | ok going to grab lunch brb | |
| 14:02:09 | openstackgerrit | do3meli proposed openstack/nova master: docs: add nova host-evacuate command to evacuate documentation https://review.openstack.org/578040 | |
| 14:06:12 | efried | sean-k-mooney: We should spell it right, though. | |
| 14:08:01 | stephenfin | efried: It's the main reason we keep jaypipes around, tbh | |
| 14:08:03 | sean-k-mooney | dambit hehe ill fix that when i get back | |
| 14:08:16 | efried | sean-k-mooney: I'll push the edit right quick. | |
| 14:08:39 | openstackgerrit | Eric Fried proposed openstack/os-vif master: add upper_constraints support https://review.openstack.org/583582 | |
| 14:08:43 | stephenfin | spellchecker extraordinaire | |
| 14:09:02 | efried | sean-k-mooney: +2 | |
| 14:09:18 | openstackgerrit | Matt Riedemann proposed openstack/os-vif master: add upper_constraints support https://review.openstack.org/583582 | |
| 14:09:33 | efried | mriedem: wtf, yo | |
| 14:10:23 | stephenfin | efried: It's alright - we'd missed anyway | |
| 14:10:59 | sean-k-mooney | mriedem: yes you are right ill fix those | |
| 14:11:13 | efried | I'm still not clear that we care about upper constraints (or constraitns or constranits) for doc builds, but okay. | |
| 14:12:12 | sean-k-mooney | oh the pti change has merged that why i missed the others | |
| 14:12:40 | jaypipes | stephenfin: glad to be of service. :) | |
| 14:12:41 | sean-k-mooney | ok got to run to shop back in 15 min ill fix it all up then | |
| 14:13:25 | stephenfin | efried: Yeah, I'm on the fence myself about them, despite suggesting it in the first place | |
| 14:13:52 | jaypipes | tssurya: done! | |
| 14:22:17 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: Transform libvirt.error notification https://review.openstack.org/484851 | |
| 14:23:48 | mriedem | tssurya: the handling a down cell changes seem to be kind of all over the place, can you rebase those into a series? | |
| 14:23:54 | mriedem | where the api changes with the microversion come at the end? | |
| 14:24:34 | mriedem | i would do in this order: object change for queued_for_delete -> online data migration -> setting queued_for_delete -> handling a down cell + microversion | |
| 14:29:11 | efried | pooja_jadhav: Are you using DiskFilter? | |
| 14:30:00 | pooja_jadhav | pooja_jadhav: I have checked nova.conf, also nova-cpu.conf.. in that in enable_filters there is no any DiskFilter is added. | |
| 14:30:52 | openstackgerrit | Matt Riedemann proposed openstack/python-novaclient master: Microversion 2.64 - Use new format policy in server group https://review.openstack.org/578261 | |
| 14:30:53 | openstackgerrit | Matt Riedemann proposed openstack/python-novaclient master: Add support for microversion 2.65 https://review.openstack.org/583632 | |
| 14:32:25 | mriedem | pooja_jadhav: which release are you on? | |
| 14:32:26 | openstackgerrit | Merged openstack/nova-specs master: Fix nits in the handling down cell spec https://review.openstack.org/581243 | |
| 14:32:39 | mriedem | master? | |
| 14:32:44 | pooja_jadhav | pooja_jadhav: Current master | |
| 14:33:07 | pooja_jadhav | mriedem: from cuurent master, fetched patch | |
| 14:33:23 | pooja_jadhav | using latest code | |
| 14:33:34 | mriedem | and you boot from volume and nova hypervisor-stats is showing local_gb_used=1? | |
| 14:33:42 | pooja_jadhav | yes | |
| 14:33:54 | mriedem | are you using devstack? | |
| 14:33:58 | pooja_jadhav | yes | |
| 14:34:23 | mriedem | and libvirt driver correct? | |
| 14:34:28 | pooja_jadhav | yes | |
| 14:34:36 | mriedem | does the flavor have ephemeral_gb? | |
| 14:35:23 | pooja_jadhav | using flavor 1 (ephemeral_gb=0, swap=0) | |
| 14:35:31 | mriedem | what is reserved_host_disk_mb set to in nova-cpu.conf? | |
| 14:36:20 | pooja_jadhav | not set any value | |
| 14:36:54 | tssurya | jaypipes: thanks | |
| 14:37:47 | mriedem | pooja_jadhav: efried: oh i see exactly why | |
| 14:38:13 | efried | mriedem: Do tell. I've been staring at this and slowly gleaning understanding, but am still out of my depth. | |
| 14:38:18 | mriedem | https://github.com/openstack/nova/blob/eb4f65a7951e921b1cd8d05713e144e72f2f254f/nova/compute/resource_tracker.py#L1455 | |
| 14:38:26 | mriedem | the usage dict doesn't know about bfv | |
| 14:38:57 | mriedem | pooja_jadhav: can you check the resource allocations for that instance to make sure DISK_GB is unset or 0? | |
| 14:39:19 | mriedem | pooja_jadhav: https://docs.openstack.org/osc-placement/latest/cli/index.html#resource-provider-allocation-show | |
| 14:39:32 | mriedem | openstack resource provider allocation show <instance uuid> | |
| 14:39:35 | tssurya | mriedem: so the what I have it now is change for qfd -> migration and setting qfd sit on this. The patch for API changes is just a POC/WIP for nova list which I am working on to include the nova show as well. | |
| 14:39:37 | mriedem | DISK_GB should either not be there or 0 | |
| 14:39:52 | tssurya | the way* | |
| 14:40:03 | pooja_jadhav | allocation_candidates not returning DISK_GB | |
| 14:40:06 | mriedem | tssurya: right i don't like how the series diverges | |
| 14:40:11 | mriedem | tssurya: it would be nice to line them all up | |
| 14:40:24 | mriedem | pooja_jadhav: you mean "openstack resource provider allocation show <instance uuid>" doesn't show DISK_GB? | |
| 14:40:29 | tssurya | mriedem: ack,I will rebase them the way you said | |
| 14:40:34 | mriedem | tssurya: thanks | |
| 14:41:01 | mriedem | pooja_jadhav: if that's the case, then it's working as designed in dansmith's change, and we just have a latent bug in the resource tracker for reporting disk uage for volume-backed instances | |
| 14:41:06 | mriedem | which isn't surprising to me | |
| 14:41:09 | pooja_jadhav | its showing {u'VCPU': 1, u'MEMORY_MB': 512} | |
| 14:41:13 | pooja_jadhav | only | |
| 14:41:14 | mriedem | and could be fixed separate from dansmith's change | |
| 14:41:20 | mriedem | efried: ^ agree? | |
| 14:41:38 | mriedem | pooja_jadhav: you could report a nova bug for that RT issue | |
| 14:41:44 | mriedem | pooja_jadhav: this specifically https://github.com/openstack/nova/blob/eb4f65a7951e921b1cd8d05713e144e72f2f254f/nova/compute/resource_tracker.py#L1455 | |
| 14:41:47 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Update queued-for-delete from the ComputeAPI during deletion/restoration https://review.openstack.org/566813 | |
| 14:42:01 | efried | mriedem: Not sure about "separate from dansmith's change". Gimme a minute here. | |
| 14:42:20 | mriedem | efried: dan's change is really about allocations in placement, | |
| 14:42:33 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: [POC] Graceful handling of nova-list when a cell is down https://review.openstack.org/567785 | |
| 14:42:36 | mriedem | there are still known issues with reporting disk usage, like if you have the DiskFilter enabled it's going to do a disk claim in the RT | |
| 14:42:48 | mriedem | we can't drop the disk claims until we drop caching scheduler | |
| 14:42:54 | mriedem | since it doesn't use placement | |
| 14:43:10 | mriedem | fixing https://github.com/openstack/nova/blob/eb4f65a7951e921b1cd8d05713e144e72f2f254f/nova/compute/resource_tracker.py#L1455 is going to be non-trivial i think | |
| 14:44:00 | mriedem | well maybe it's not that bad, we have an instance uuid in scope when it's used and from the instance we can determine if it's volume-backed or not | |
| 14:47:10 | efried | mriedem: How do I access the compute node object from the compute service thingy (self.compute1)? | |
| 14:49:13 | pooja_jadhav | mriedem: you mean to say 'root_gb' is taking directly from flavor.root_gb. whether we create instance from volume or not? correct? | |
| 14:49:17 | mriedem | efried: in a functional test? | |
| 14:49:23 | mriedem | pooja_jadhav: correct | |
| 14:49:31 | efried | mriedem: yeah | |
| 14:49:36 | mriedem | efried: self.compute1 is the service record, | |
| 14:50:36 | mriedem | from that you can get the manager and host i think, and from that you can get the compute node record by hostname and then get it's uuid | |
| 14:50:42 | mriedem | self.compute1.manager.host | |
| 14:51:15 | mriedem | that's how ComputeManager._get_compute_nodes_in_db works anyway | |
| 14:51:16 | efried | mriedem: That's just a string. | |
| 14:52:04 | mriedem | right to lookup the compute node | |
| 14:52:14 | pooja_jadhav | mriedem: in that case, we need to check, if the instance is volume_backed then make root_gb is to 0. | |
| 14:52:24 | mriedem | i guess you could use self.compute1.manager._resource_tracker.compute_nodes | |