Closed
Bug 1074199
Opened 11 years ago
Closed 11 years ago
Add clear_cache command to the production update script
Categories
(Tree Management :: Treeherder, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: mdoglio, Assigned: mdoglio)
References
Details
Attachments
(1 file)
See bug 1074077 and bug 1072291
| Assignee | ||
Updated•11 years ago
|
Assignee: nobody → mdoglio
Status: NEW → ASSIGNED
| Assignee | ||
Comment 1•11 years ago
|
||
Attachment #8496889 -
Flags: review?(jeads)
| Assignee | ||
Updated•11 years ago
|
Status: ASSIGNED → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
Comment 2•11 years ago
|
||
Commits pushed to master at https://github.com/mozilla/treeherder-service
https://github.com/mozilla/treeherder-service/commit/703bf33297d6e75cf78c76f62580834a9a55cfac
Bug 1074199 - add clear_cache command to the update script
https://github.com/mozilla/treeherder-service/commit/f1bc715cfd81d659226bb7308ad2b9a8fbd304b1
Merge pull request #231 from mozilla/add-clear-cache-command
Bug 1074199 - add clear_cache command to the update script
Updated•11 years ago
|
Attachment #8496889 -
Flags: review?(jeads) → review+
Comment 3•11 years ago
|
||
Is this is clearing out the cached top_revisions for each repo, right? (as well as anything else in the cache)
If so, I think this is dangerous to have in update.py, since if there are infra issues at the time of the deploy (eg like there were last night), then we fall back to json-pushes with no params, which only fetches the last 10 pushes. Which if there was a backlog in terms of pushlog ingestion, might mean we miss pushes older than #10.
Flags: needinfo?(jeads)
Comment 4•11 years ago
|
||
(In reply to Ed Morley [:edmorley] from comment #3)
> Is this is clearing out the cached top_revisions for each repo, right? (as
> well as anything else in the cache)
>
> If so, I think this is dangerous to have in update.py, since if there are
> infra issues at the time of the deploy (eg like there were last night), then
> we fall back to json-pushes with no params, which only fetches the last 10
> pushes. Which if there was a backlog in terms of pushlog ingestion, might
> mean we miss pushes older than #10.
Bug 1076792 will reduce the risk slightly, but I'm wondering whether we should still make the cache clear manual-only?
You need to log in
before you can comment on or make changes to this bug.
Description
•