Bug 211248 - results.webkit.org: Dropped connections when uploading archives shouldn't be fatal
Summary: results.webkit.org: Dropped connections when uploading archives shouldn't be ...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Jonathan Bedard
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2020-04-30 12:25 PDT by Jonathan Bedard
Modified: 2020-05-01 08:35 PDT (History)
5 users (show)

See Also:


Attachments
Patch (1.56 KB, patch)
2020-04-30 12:28 PDT, Jonathan Bedard
no flags Details | Formatted Diff | Diff
Patch (3.04 KB, patch)
2020-04-30 12:37 PDT, Jonathan Bedard
no flags Details | Formatted Diff | Diff
Patch (3.77 KB, patch)
2020-04-30 20:08 PDT, Jonathan Bedard
no flags Details | Formatted Diff | Diff
Patch (3.97 KB, patch)
2020-05-01 07:46 PDT, Jonathan Bedard
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jonathan Bedard 2020-04-30 12:25:22 PDT
Result archives for layout tests can be unreasonably large, especially if there are many crashes. Failing to upload these archives should not be fatal.
Comment 1 Jonathan Bedard 2020-04-30 12:28:26 PDT
Created attachment 398081 [details]
Patch
Comment 3 Jonathan Bedard 2020-04-30 12:37:42 PDT
Created attachment 398083 [details]
Patch
Comment 4 Aakash Jain 2020-04-30 13:20:41 PDT
Comment on attachment 398083 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=398083&action=review

> Tools/Scripts/webkitpy/results/upload.py:229
> +            log_line_func(' ' * 4 + 'Failed to upload test archive to {}, results server dropped connection, likely due to archive size.'.format(hostname))

Can we also print the archive size here?

> Tools/Scripts/webkitpy/results/upload.py:231
> +            return True

What happens is server is down or unreachable? Do we still use this code path or print a different error?
Comment 5 Jonathan Bedard 2020-04-30 14:03:14 PDT
Comment on attachment 398083 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=398083&action=review

>> Tools/Scripts/webkitpy/results/upload.py:231
>> +            return True
> 
> What happens is server is down or unreachable? Do we still use this code path or print a different error?

It does, but it’s unlikely we would ever get into that circumstance because the result upload would have also failed.
Comment 6 Jonathan Bedard 2020-04-30 14:03:15 PDT
Comment on attachment 398083 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=398083&action=review

>> Tools/Scripts/webkitpy/results/upload.py:231
>> +            return True
> 
> What happens is server is down or unreachable? Do we still use this code path or print a different error?

It does, but it’s unlikely we would ever get into that circumstance because the result upload would have also failed.
Comment 7 Jonathan Bedard 2020-04-30 20:08:48 PDT
Created attachment 398150 [details]
Patch
Comment 8 Jonathan Bedard 2020-05-01 07:46:27 PDT
Created attachment 398185 [details]
Patch
Comment 9 EWS 2020-05-01 08:34:37 PDT
Committed r261000: <https://trac.webkit.org/changeset/261000>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 398185 [details].
Comment 10 Radar WebKit Bug Importer 2020-05-01 08:35:15 PDT
<rdar://problem/62728138>