1270211 - Reprocessing request for ipc_message_error

Status: RESOLVED FIXED

(Socorro :: General, task)

Description

[Benjamin Smedberg]

Once bug 1269817 is deployed, please reprocess the following crashes:

* any crash with the ipc_message_error annotation
* 27-april through now

Which should be equivalent to this supersearch:
https://crash-stats.mozilla.com/search/?ipc_channel_error=!__null__&date=%3E%3D2016-04-27&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#crash-reports

Comment 1

[Peter Bengtsson [:peterbe]]

https://bugzilla.mozilla.org/show_bug.cgi?id=1269817 IS deployed to prod as of Wednesday May 4th. Can you confirm that it's working as expected?

Comment 2

[Benjamin Smedberg]

I confirm that we're getting the correct new data.

Comment 3

[Peter Bengtsson [:peterbe]]

I have a script ready. It's kinda brutal but it should work. 
https://gist.github.com/a33bbf20b77ca55ee07ed0ebce5a2091
You run it like this: `python download_all_crash_ids_bug_1270211.py :avalidauthtoken:` and it generates a .sql file can you can load into the production database. 

At the time of writing, there are 9439 unique crash IDs for that SuperSearch query. It'll be slightly more the actual time we run it but I guess if the code is in production it's not a huge harm to do some twice that already have it in them.

Comment 4

[Peter Bengtsson [:peterbe]]

Can you look at https://crash-stats.allizom.org/search/?ipc_channel_error=!__null__&date=%3E%3D2016-04-27&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#crash-reports and click on some of them to see that the reprocessing worked?
I started the reprocessing there 10min ago but it was only ~1,000 crash_ids (which makes sense since we throttle 10% of crashes over to stage). 

I randomly clicked on one (https://crash-stats.allizom.org/report/index/f042379a-214f-4232-b299-6c5dc2160429) and it says "Signature replaced with an IPC Channel Error, was: "libsystem_kernel.dylib@0x16db6" in the Processor Notes. 

If that's satisfactory (on stage), I'll proceed equally with prod after lunch.

Comment 5

[Benjamin Smedberg]

Yes reprocessing works, according to those links. (Please use NEEDINFO, I wouldn't have seen your request if I hadn't been sorting bugmail).

Comment 6

[Peter Bengtsson [:peterbe]]

This is now being reprocessed in prod. I don't think it should take particularly long. Perhaps give it an hour at least and try to your query again and spot check. I'll let you resolve the bug if you think it worked.

Comment 7

[Benjamin Smedberg]

Verified. Thank you.