yahoo-eng-team team mailing list archive
-
yahoo-eng-team team
-
Mailing list archive
-
Message #19757
[Bug 1283987] Re: Query Deadlock when creating >200 servers at once in sqlalchemy
** Project changed: oslo-incubator => oslo.db
--
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/1283987
Title:
Query Deadlock when creating >200 servers at once in sqlalchemy
Status in OpenStack Compute (Nova):
In Progress
Status in Oslo Database library:
In Progress
Bug description:
Query Deadlock when creating >200 servers at once in sqlalchemy.
--------
This bug occurred when I test this bug:
https://bugs.launchpad.net/nova/+bug/1270725
The original info is logged here:
http://paste.openstack.org/show/61534/
--------------
After checking the error-log, we can notice that the deadlock function
is 'all()' in sqlalchemy framework.
Previously, we use '@retry_on_dead_lock' function to retry requests when deadlock occurs.
But it's only available for session deadlock(query/flush/execute). It
doesn't cover some 'Query' actions in sqlalchemy.
So, we need to add the same protction for 'all()' in sqlalchemy.
To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1283987/+subscriptions
References