[a10n] update raven to avoid encoding problems

RESOLVED FIXED

Status

Webtools
Elmo
RESOLVED FIXED
4 years ago
4 years ago

People

(Reporter: Pike, Assigned: rhelmer)

Tracking

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

(Reporter)

Description

4 years ago
The a10n logs are pretty full of encoding problems in raven,

    return codecs.utf_8_decode(input, errors, True)
UnicodeDecodeError: 'utf8' codec can't decode byte 0xd9 in position 3: invalid continuation byte
'utf8' codec can't decode byte 0xf9 in position 0: invalid start byte
Traceback (most recent call last):
  File "/data/src/a10n/env/lib/python2.6/site-packages/raven/utils/serializer/manager.py", line 67, in transform
    return serializer.serialize(value, **kwargs)
  File "/data/src/a10n/env/lib/python2.6/site-packages/raven/utils/serializer/base.py", line 117, in serialize
    value.decode('utf-8').encode('utf-8')[:string_max_length],)
  File "/data/src/a10n/env/lib64/python2.6/encodings/utf_8.py", line 16, in decode


going around and around.

Getting a bug on file to investigate, not really sure at this point where to start.
(Reporter)

Comment 1

4 years ago
This looks very much like https://github.com/getsentry/raven-python/issues/292. There are actually a few encoding-related open issues on raven :-/, but none strictly this.
(Reporter)

Comment 2

4 years ago
Seems that https://github.com/getsentry/raven-python/commit/d3fde3fc35a7372f72b6ca4c34ee32aeab4e2771 is on our stack trace, and now different.

We should try to update Raven to something current, and see if that retriggers.

Rob, do you want to give this a stab? I'd go bold and try 4.0.3.
(Reporter)

Updated

4 years ago
Summary: [a10n] investigate raven encoding problems → [a10n] update raven to avoid encoding problems
(Assignee)

Updated

4 years ago
Assignee: nobody → rhelmer
Status: NEW → ASSIGNED
(Assignee)

Comment 3

4 years ago
Looks like this is 3.3.7, both in requirements.txt and in the virtualenv "env" that the a10n server uses:

>>> import raven
>>> raven.VERSION
'3.3.7'
(Assignee)

Comment 4

4 years ago
PR: https://github.com/Pike/a10n/pull/5

I've upgraded this on the a10n server and restarted too.. no errors yet though! Maybe we should make one happen to test.
(Reporter)

Comment 5

4 years ago
Pulled this in to master.

FYI, you can just paste 'https://github.com/Pike/a10n/pull/5' as an attachment, and bugzilla will treat it like a URL to a patch. That's what the gaia team uses extensively. No further options needed.
Status: ASSIGNED → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
(Reporter)

Comment 6

4 years ago
We need to back this out, sadly, the current raven uses a protocol version that's not supported by our sentry install. I think we need to go back to a 3.6.1, seems that's the latest version supporting protocol version 2.0 (latest is 4)
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
(Reporter)

Comment 7

4 years ago
Created attachment 8419394 [details] [diff] [review]
roll back raven to 3.6.1
Attachment #8419394 - Flags: review?(rhelmer)
(Assignee)

Updated

4 years ago
Attachment #8419394 - Flags: review?(rhelmer) → review+
(Assignee)

Comment 8

4 years ago
Opened a PR for this, I can update the a10n when it's merged:
https://github.com/Pike/a10n/pull/6
(Reporter)

Comment 9

4 years ago
As I mentioned in the PR, I frown upon merge commits. As this is really just a hack to work around you not being able to push to my repo, I fixed that instead.
(Assignee)

Comment 10

4 years ago
* pushed to pike's repo
* pulled latest git on a10n server
* "pip install -r requirements.txt" into the virtualenv
* restarted supervisord

Looks like everything is running OK in the local logs, will check errormill periodically and verify bug when I see errors come in.
Status: REOPENED → RESOLVED
Last Resolved: 4 years ago4 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.