New issue
Advanced search Search tips

Issue 678396 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Jan 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: All
Pri: 2
Type: Bug



Sign in to add a comment

Swarming should be resilient to isolated_cache being purged

Project Member Reported by aga...@chromium.org, Jan 4 2017

Issue description

What steps will reproduce the problem?
1. Have a bot with low disk space (e.g. https://buganizer.corp.google.com/issues/34049848)
2. Delete items from isolated_cache because it is large (see comment 2)

What is the expected result?
3. Isolate re-downloads items if it needs them again

What happens instead of that?
3. It breaks (see comment 3)

It's nice that isolate tries to clean up this directory itself, and apparently we're already working on tweaking the threshold (see comment 4), but isolated_cache is a cache. Items being deleted from it should not lead to fatal failures.
 
Owner: vadimsh@chromium.org
Status: Assigned (was: Available)
Looks like a regression. It used to be resilent to cache corruption: https://github.com/luci/luci-py/blob/master/client/isolateserver.py#L771

This function should catch ENOENT and return False (instead of raising OSError), it will be equivalent to a cache miss: https://github.com/luci/luci-py/blob/master/client/isolateserver.py#L348 

I'll make this change.
Status: Fixed (was: Assigned)
This has been deployed.

Sign in to add a comment