596766 - Disable crons and clear out bad aggregate data

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: WebOps: Other, task)

Description

[Austin King [:ozten]]

Last night's release has corrupted the crash signatures in the reports and top crashers by signature view for window of time.

Metrics is rebuilding out backend view which will fix these entries in the reports table.

We need to disable the crons and then remove the bad entries.

Steps:
1) remove startTopCrashersBySignature.py from the crontab
2) Excute the following SQL
DELETE FROM top_crashes_by_signature WHERE window_end > '2010-09-14 12:00';

Comment 1

[Aravind Gottipati [:aravind]]

In staging

breakpad=> DELETE FROM top_crashes_by_signature WHERE window_end > '2010-09-14 12:00';
DELETE 0
breakpad=> 

Is that expected, okay to run it in production?

Comment 2

[Austin King [:ozten]]

(In reply to comment #1)
Okay. Please run in prod.

Comment 3

[Aravind Gottipati [:aravind]]

breakpad=> DELETE FROM top_crashes_by_signature WHERE window_end > '2010-09-14 12:00';
DELETE 107702
breakpad=> \q


Cron job commented out.

Comment 4

[Robert Helmer [:rhelmer]]

Seeing this right now in prod and stage (bug 596881 comment 3 and bug 596881 comment 6).

Do we need to do something like this again?

Comment 5

[Robert Helmer [:rhelmer]]

(In reply to comment #4)
> Seeing this right now in prod and stage (bug 596881 comment 3 and bug 596881
> comment 6).
> 
> Do we need to do something like this again?

That bug is about top crashers by domain, BTW.