Merge lp:~jml/txpkgme/long-check into lp:txpkgme
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | Jonathan Lange | ||||
Approved revision: | 31 | ||||
Merged at revision: | 30 | ||||
Proposed branch: | lp:~jml/txpkgme/long-check | ||||
Merge into: | lp:txpkgme | ||||
Diff against target: |
144 lines (+66/-5) 2 files modified
txpkgme/submitfromdisk.py (+20/-5) txpkgme/tests/test_submitfromdisk.py (+46/-0) |
||||
To merge this branch: | bzr merge lp:~jml/txpkgme/long-check | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Jonathan Lange (community) | Approve | ||
Review via email: mp+120198@code.launchpad.net |
Commit message
Add --shutdown-delay option to delay shutting down the webserver.
Description of the change
We're getting error messages in our celery log and in our OOPS reports saying
"Connection Refused". Really, these errors are mostly due to a buggy client,
submit-
providing a response.
It would be nice to prevent this properly. However, until then, to silence
these warnings and to make debugging nagios failures easier, let's just add
a delay between the webserver sending a response and shutting itself down.
We'll need a txpkgme rollout and an update to the cronjob that feeds the
nagios check (adding '--shutdown-delay 5') before this kicks in.
I give this a tentative +1 - tentative because its friday evening and also because my twisted foo is weak. A second opinion is probably good, but otoh its compact enough for low risk.