Merge lp://staging/~ev/daisy/consistency-problems into lp://staging/daisy
Proposed by
Evan
| Status: | Merged |
|---|---|
| Merged at revision: | 191 |
| Proposed branch: | lp://staging/~ev/daisy/consistency-problems |
| Merge into: | lp://staging/daisy |
| Diff against target: |
40 lines (+6/-3) 1 file modified
process_core.py (+6/-3) |
| To merge this branch: | bzr merge lp://staging/~ev/daisy/consistency-problems |
| Related bugs: |
| Reviewer | Review Type | Date Requested | Status |
|---|---|---|---|
| Daisy Pluckers | Pending | ||
|
Review via email:
|
|||
Description of the change
This should fix the consistency problems we're seeing on the retracers. They're currently getting NotFoundExceptions on the initial lookup of the OOPS ID in the OOPS CF, despite it definitely being there:
https:/
I've confirmed this to work in production.
Preview Diff
[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
From #webops, 20121205:
[16:26:59] <ev> it is indeed an eventual consistency problem, I'm just not sure why we're only seeing it now
[16:34:53] <thedac> ev: We added the new nodes last week. Has this been since then or just recently?
[16:35:30] <ev> ah, maybe that's it then
[16:35:40] <ev> thedac: because we were at RF=3 with three nodes before, right?
[16:35:47] <ev> and now we're at RF=3 with 6 nodes
[16:35:51] <thedac> right
[16:36:05] <ev> so I imagine we'll need to be reading at quorum in other places as well soon :)