No, my hypothesis doesn't look correct. We don't see anything that
would make prepare() jump into the reactor thread.
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to maas in Ubuntu.
https://bugs.launchpad.net/bugs/1308069
Title:
New hypothesis: the code in DatabaseLock opens, and closes, a cursor for
each locking/unlocking command. Do we actually know that these cursors
will be in the same database session? If the command failed, do we know
that we would see an error?
--
You received this bug notification because you
** Branch linked: lp:~allenap/maas/database-locks-revisited
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to maas in Ubuntu.
https://bugs.launchpad.net/bugs/1308069
Title:
test_prepare_holds_startup_lock() fails spuriously
To manage
** Branch unlinked: lp:~allenap/maas/database-locks-revisited
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to maas in Ubuntu.
https://bugs.launchpad.net/bugs/1308069
Title:
test_prepare_holds_startup_lock() fails spuriously
To
It happened again. I suspect that it may be a matter of ordering of
decorators: RegionAdvertisingService.prepare is decorated as
@synchronous, and *then* as taking two locks.
Given decorators' wrapping behaviour, which reverses the order of
entrance, I understand that to mean: grab these two
** Also affects: maas (Ubuntu Trusty)
Importance: Undecided
Status: New
** Changed in: maas (Ubuntu)
Status: New = Fix Released
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to maas in Ubuntu.
** Also affects: maas (Ubuntu)
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to maas in Ubuntu.
https://bugs.launchpad.net/bugs/1308069
Title:
test_prepare_holds_startup_lock() fails
** Description changed:
+ [Test Case]
+ No test case; the code that's being patched is only a test and does not
actually appear in the package.
+
+
+ [Description of the problem]
+
This happened when trying to land a documentation-only branch: