| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-04-18 | |||
| 07:31:58 | openstackgerrit | Boxiang Zhu proposed openstack/nova master: Add host and hypervisor_hostname flag to create server https://review.openstack.org/645520 | |
| 08:01:28 | openstackgerrit | Hamdy Khader proposed openstack/nova master: Do not perform port update in case of baremetal instance. https://review.openstack.org/649345 | |
| 08:15:39 | openstackgerrit | Takashi NATSUME proposed openstack/python-novaclient master: Fix a description for config_drive parameter https://review.openstack.org/653683 | |
| 09:04:47 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Plumbing for locking an instance with reason https://review.openstack.org/653691 | |
| 09:04:47 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: [WIP] Support adding the reason behind a server lock https://review.openstack.org/648662 | |
| 09:14:47 | openstackgerrit | Boxiang Zhu proposed openstack/nova master: Make evacuation respects anti-affinity rule https://review.openstack.org/649963 | |
| 09:16:59 | openstackgerrit | Boxiang Zhu proposed openstack/nova master: Fix live migration break group policy simultaneously https://review.openstack.org/651969 | |
| 09:19:24 | plestang | Hi all! | |
| 09:19:49 | plestang | I have a question regarding nova-conductor | |
| 09:20:29 | plestang | regurlarly compute hosts are requesting for unconfirmed migration by dest compute | |
| 09:21:07 | plestang | The function called is here: https://github.com/openstack/nova/blob/03322bb517925a9f5a04ebdb41c3fd31e7962440/nova/db/sqlalchemy/api.py#L4313 | |
| 09:21:23 | plestang | the read_deleted params is set to yes | |
| 09:21:30 | plestang | in model_query | |
| 09:21:50 | plestang | it generates some load on our infra | |
| 09:22:08 | plestang | because we have around 300000 lines in that table | |
| 09:22:59 | plestang | so all compute hosts of our infra reads regularly all the database | |
| 09:24:45 | plestang | when I explain the request | |
| 09:25:16 | plestang | +------+---------------+------------+--------+-----------------+--------+-----------+--------+--------+-------------+ | |
| 09:25:18 | plestang | | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | | |
| 09:25:20 | plestang | |------+---------------+------------+--------+-----------------+--------+-----------+--------+--------+-------------| | |
| 09:25:22 | plestang | | 1 | SIMPLE | migrations | ALL | <null> | <null> | <null> | <null> | 261177 | Using where | | |
| 09:25:24 | plestang | +------+---------------+------------+--------+-----------------+--------+-----------+--------+--------+-------------+ | |
| 09:25:57 | plestang | I was asking if it were a good idea to set the read_deleted to no instead of yes | |
| 09:26:40 | plestang | it would make the MariaDB query plan use one of the index that use the delete column | |
| 09:27:52 | plestang | if I add "and deleted=0" the query plan becomes | |
| 09:27:54 | plestang | +------+---------------+------------+--------+---------------------------------------------------------------------------------+-----------------------------------------+-----------+-------+--------+------------------------------------+ | |
| 09:27:57 | plestang | | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | | |
| 09:27:59 | plestang | |------+---------------+------------+--------+---------------------------------------------------------------------------------+-----------------------------------------+-----------+-------+--------+------------------------------------| | |
| 09:28:01 | plestang | | 1 | SIMPLE | migrations | ref | migrations_by_host_nodes_and_status_idx,migrations_instance_uuid_and_status_idx | migrations_instance_uuid_and_status_idx | 5 | const | 23610 | Using index condition; Using where | | |
| 09:28:03 | plestang | +------+---------------+------------+--------+---------------------------------------------------------------------------------+-----------------------------------------+-----------+-------+--------+------------------------------------+ | |
| 09:28:54 | plestang | in other word the question is what is the idea behind requesting finished migration in deleted status for compute host | |
| 10:35:00 | aspiers | plestang: you might have more luck when the Americans come online | |
| 10:35:15 | aspiers | kashyap: any thoughts on how to split the getDomainCapabilities() patch up? | |
| 10:35:34 | kashyap | aspiers: Hiya, I want to spend some careful time on it | |
| 10:35:58 | kashyap | aspiers: Right now, I'm chugging through a few things that I "need" to finish | |
| 10:36:03 | aspiers | I guess I could manufacture something to test which isn't ever used in the real world | |
| 10:36:07 | kashyap | aspiers: Rest assured, since I have a "vested interest" in it | |
| 10:36:08 | aspiers | OK np | |
| 10:36:14 | aspiers | Sure, I noticed ;-) | |
| 10:36:19 | kashyap | Hehe | |
| 10:36:22 | aspiers | In the mean time I guess I can rebase at least | |
| 10:36:30 | kashyap | Let me quickly re-read it, actually | |
| 10:36:44 | aspiers | Need to fix the merge conflicts | |
| 10:47:32 | openstackgerrit | Ivaylo Mitev proposed openstack/nova master: VMware: Attach volumes using adapter type from instance https://review.openstack.org/616599 | |
| 10:51:33 | stephenfin | dansmith: When you're about, this is probably worth you looking at https://review.openstack.org/#/c/651296/ | |
| 11:00:29 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: autospecs all the mock.patch usages https://review.openstack.org/470775 | |
| 11:00:48 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hyper-v: autospec classes before they are instantiated https://review.openstack.org/342211 | |
| 11:16:17 | kashyap | aspiers: Responded on the change. I can understand if my response is not fully satisfying... | |
| 11:17:04 | aspiers | kashyap: thanks! | |
| 11:17:51 | kashyap | First see if I make sense :-) Free to correct me if I didn't consider something else. | |
| 11:19:04 | aspiers | sure | |
| 11:41:40 | openstackgerrit | Hamdy Khader proposed openstack/nova master: Do not perform port update in case of baremetal instance. https://review.openstack.org/649345 | |
| 12:05:51 | plestang | aspiers: ok thx waiting a couple of hours | |
| 12:12:40 | openstackgerrit | Lee Yarwood proposed openstack/nova master: fixtures: Return a mocked class instead of method within fake_imagebackend https://review.openstack.org/619804 | |
| 12:43:32 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposes NUMA affinity for vGPUs https://review.openstack.org/650963 | |
| 12:44:03 | bauzas | dansmith: efried: thanks for commenting the vGPU affinity spec, I made it way simplier ^ | |
| 12:45:14 | bauzas | efried: about your comment on the new field name, I just wanted to make this field as arbitrary as possible, since other features could want to affinitize it too (think of... network ports :) heh gibi) | |
| 13:02:37 | efried | bauzas: But you're already calling it "devices". The part that makes it clearer is the "rp_uuids" bit. Otherwise the only thing you have to go on is the fact that it's a UUIDField. Are you saying it could include UUIDs *other* than RP UUIDs? How would the code know which is which?? | |
| 13:03:07 | bauzas | ok, we can call it rp_uuids if so | |
| 13:03:18 | bauzas | or related_rp_uuids | |
| 13:03:27 | bauzas | FWIW, I'm very bad at naming things but my children | |
| 13:03:28 | sean-k-mooney | bauzas: ya i think rp_uuids is good | |
| 13:03:40 | sean-k-mooney | -1 to related_rp_uuids | |
| 13:03:50 | sean-k-mooney | too much typeing | |
| 13:05:01 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposes NUMA affinity for vGPUs https://review.openstack.org/650963 | |
| 13:05:02 | bauzas | sean-k-mooney: efried: my pleasure ^ | |
| 13:06:13 | efried | bauzas: So now does it include the UUID of the NUMA cell RP? | |
| 13:06:44 | sean-k-mooney | no | |
| 13:07:14 | sean-k-mooney | maybe child_rps woould be better... | |
| 13:07:30 | sean-k-mooney | actully it could | |
| 13:07:32 | efried | Yeah, good idea, child_rp_uuids | |
| 13:07:47 | sean-k-mooney | we coudl do it either way | |
| 13:08:04 | sean-k-mooney | if it doen not include it we shoudl have a field for the numa node RP | |
| 13:08:21 | efried | I don't love including the NUMA cell RP UUID, because how would you tell which one that is. But yes, this ^ | |
| 13:08:27 | sean-k-mooney | but this spec does not depend on have numa nodes as RPs in placment | |
| 13:08:41 | efried | then make the field nullable | |
| 13:09:03 | efried | or set it to the compute RP UUID. | |
| 13:09:21 | efried | um, not that second thing - because there can still be more than one, eh? | |
| 13:10:02 | sean-k-mooney | so in this spec i think have it as just child_rp_uuids makes sense | |
| 13:10:25 | sean-k-mooney | and if we also do the numa in placement spec add a rp_uuid filed for the numa node RP | |
| 13:11:41 | efried | - Name the new ListOfUUIDField `child_rp_uuids` and it includes all & only UUIDs of child RPs like PGPUs, NICs, etc. | |
| 13:11:41 | efried | - Add rp_uuid (the UUID of the NUMA cell) as a nullable UUIDField, only set if the NUMA cell is represented as a nested RP | |
| 13:11:41 | efried | So summarizing: | |
| 13:12:06 | sean-k-mooney | +1 | |
| 13:12:18 | sean-k-mooney | bauzas: what do you think ^ | |
| 13:14:27 | bauzas | efried: I don't want the vGPU affinity spec to tell about any NUMA node being a nested RP | |
| 13:14:31 | bauzas | it's a different spec | |
| 13:14:58 | bauzas | for the moment, pGPUs are a specific RP, sure, but NUMA nodes are just using the root RP | |
| 13:15:11 | sean-k-mooney | bauzas: well you could leave adding the the nullable rp_uuid fiedl to the other spec | |
| 13:15:58 | bauzas | sean-k-mooney: maybe, I'm not sure we need it | |
| 13:16:13 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposes NUMA affinity for vGPUs https://review.openstack.org/650963 | |
| 13:16:14 | bauzas | bikeshed update ^ | |
| 13:17:00 | sean-k-mooney | :) that looks fine too me | |
| 13:17:55 | sean-k-mooney | the other thing i wanted to point out is if we did not have the numa-as-nested-rp spec then this is enought to do numa with placement in general | |
| 13:18:51 | openstack | Launchpad bug 1823251 in OpenStack Compute (nova) "Spike in TestNovaMigrationsMySQL.test_walk_versions/test_innodb_tables failures since April 1 2019 on limestone-regionone" [High,Confirmed] | |
| 13:18:51 | efried | mdbooth: https://bugs.launchpad.net/nova/+bug/1823251 <== have you had a chance to look into this? (weird db race brought up two nova meetings ago) | |
| 13:18:52 | sean-k-mooney | take sriov as an example if we just model the pfs as RPs of the compute node and add the rps to the child_rp_uuids then it will work | |
| 13:19:21 | sean-k-mooney | well we will need to do a few tweeks but the datamdoel will support it | |
| 13:20:05 | sean-k-mooney | so if we dont make progress on the richer query syntax this cycle then we can atleast model the capasity with placement and use the numa toplogy filter to track the numa locality | |
| 13:21:41 | efried | sean-k-mooney: Can you give me a rundown of the reviews you have open for CI-ish stuff? | |