yahoo-eng-team team mailing list archive

[Launchpad Bug Tracker <1828535@xxxxxxxxxxxxxxxxxx>]

[Expired for OpenStack Compute (nova) because there has been no activity
for 60 days.]

** Changed in: nova
       Status: Incomplete => Expired

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1828535

Title:
  [Perf] Archival job destroying records from api_database in bulk
  causes orphaning of records

Status in OpenStack Compute (nova):
  Expired

Bug description:
  When the nova-manage db archive_deleted rows cron jobs from several
  cells run in parallel (even if they are randomized during the day if
  there are a lot of cells this may happen) they all try to destroy in
  bulk the instance_mappings/instance_group_members/request_specs from
  the nova_api database which means each cell gets a lock on the
  api_database during which another cell would not be able to reap the
  records from the nova_api database. We have a patch for making this
  command multi-cells aware (https://review.opendev.org/#/c/507486/),
  however wondering if destroying the records one after the other in
  loop is better than destroying them in bulk ?

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1828535/+subscriptions