Properly wait for resolve snapshots work to finish instead of failing and retrying #1750
Comments
|
This seems to be the "expected" behavior. Isn't the whole point of |
|
Oh I see... in this case I suspect it's just that the condition It's a bit of an abuse of failures (failures imply somebody should be looking at a problem), couldn't it just schedule to retry a little later with standard callbacks like we do elsewhere? |
|
I don't know the rationale behind implementing it this way, but yes, having it fail is quite confusing. I do think that scheduling it to run more (i.e. return |
|
Actually the problem is worst than that: Glancing at the code, I think there is a race condition:
So we need to fix 2 things as far as I can tell:
|
|
Regarding race condition, it seems like a fix would be removing redundant |
|
yes, |
|
ok, i can look into this one. |
|
I think the delay issue will be handled with the new design in #1755 |
marta-lokhova
changed the title
|
Resolved by #2063 |
Issue type
bug
Version found
10.0.0rc1
Issue description
When a snapshot gets published, it fails to properly take the snapshot, causing a delay in publishing.
Actual result
Delay is systematic, logs show that it marks the work as failed (and retries) without trying to do anything (at least no processes are started).
Expected result
No error when publishing.
Supporting files
The text was updated successfully, but these errors were encountered: