Fix flytestdlib's stowStore.List for google cloud storage#6098
Conversation
Signed-off-by: Fabio Graetz <fabiograetz@googlemail.com>
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #6098 +/- ##
=======================================
Coverage 37.10% 37.10%
=======================================
Files 1318 1318
Lines 132403 132413 +10
=======================================
+ Hits 49122 49132 +10
- Misses 79008 79009 +1
+ Partials 4273 4272 -1
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| u := url.URL{ | ||
| Scheme: protocol, | ||
| Host: containerName, | ||
| Path: item.Name(), | ||
| } | ||
| results[index] = DataReference(u.String()) |
There was a problem hiding this comment.
Can we fix this in our fork of stow? It looks like we change the scheme here, which is used in gcs implementation of convertToStowItem.
There was a problem hiding this comment.
Will do, overlooked we use a fork of stow 👌
Closing this PR in favour of flyteorg/stow#17 @eapolinario |
Why are the changes needed?
For RFC #5598, flytepropeller was given the ability to list error files in the so-called raw output prefix bucket of an execution with the goal of identifying which worker pod in a failed distributed task experienced the first error.
For this purpose, the
StowStorein flytestdlib was given aListfunction.For google cloud storage buckets, the listing and subsequent access of the error files currently does not work: When listing a bucket
gs://some-bucket/..., one receives items in the formgoogle://storage.googleapis.com/download/storage/v1/b/some-bucket/...which then cannot be found by the stow store for thegs://prefix.This PR fixes this.
What changes were proposed in this pull request?
Don't use
item.URL()but construct the url from the protocol, the container name (bucket name), and the item name.How was this patch tested?
Tested in a GCP deployment and adapted unit test.
Related PRs