Closed Bug 822432 Opened 13 years ago Closed 13 years ago

reprocess elfhack-broken B2G crashes

Categories

(Socorro :: Data request, task)

Product:
Socorro
Component:
Data request
Platform:
All
Gonk (Firefox OS)
Type:
task
Priority:
Not set
Severity:
blocker

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: benjamin, Assigned: rhelmer)

References

Details

Attachments

(1 file, 1 obsolete file)

B2G minidump fixer
4.61 KB, text/x-csrc
Details
+++ This bug was initially created as a clone of Bug #637680 +++ Bug 637680 has all the programs, scripts and information, but basically elfhack is causing all the B2G crash dumps to be meaningless. We want to go back through all the B2G crashes for the last 3 weeks and reprocess them after running this dump script. And we may need to keep running this for another few days until we find a better client-side solution. After the reprocessing, Kairo should be able to regenerate his reports with real topcrash data.
Assignee: nobody → rhelmer
Blocks: 822476
This is actually all still in place, we simply need to adjust the SQL query for the appropriate broken products/versions. Should this just be B2G, any version, for the time being?
Component: General → Data request
Summary: Get top crashers for Firefox and Fennec where crash-stats are broken (linux, android) → reprocess elfhack-broken B2G crashes
lars r? https://github.com/mozilla/socorro/pull/1015
Status: NEW → ASSIGNED
Flags: needinfo?(lars)
(In reply to Robert Helmer [:rhelmer] from comment #1) > Should this just be B2G, any version, for the time being? I'd say all B2G where the OS comes in as "Android".
Commits pushed to master at https://github.com/mozilla/socorro https://github.com/mozilla/socorro/commit/937e714c246e48d9e4b7d010758cbb680ab56acf bug 822432 - add support for b2g to broken dump fixer https://github.com/mozilla/socorro/commit/70dfc468a56a59cc3eb3455fa8910333ff8669ba Merge pull request #1015 from rhelmer/bug822432-fix-broken-b2g-crashes bug 822432 - add support for b2g to broken dump fixer
I'd like to uplift this to stage branch now and ship w/ Socorro 31 on Wednesday, December 19 2012. I can backfill the last three weeks manually tomorrow, and then again on Wednesday (or sooner if anyone needs it) at which point the cron job should start keeping it up to date. Any objections?
Flags: needinfo?(laura)
Target Milestone: --- → 31
(In reply to Robert Helmer [:rhelmer] from comment #5) > I'd like to uplift this to stage branch now and ship w/ Socorro 31 on > Wednesday, December 19 2012. > > I can backfill the last three weeks manually tomorrow, and then again on > Wednesday (or sooner if anyone needs it) at which point the cron job should > start keeping it up to date. > > Any objections? This sounds good. I think we are going to have to do some tree-juggling though: I'll send email.
Flags: needinfo?(laura)
Blocks: 822584
I was trying to verify that the fixup tool would actually do something useful, and currently it doesn't appear to.
(In reply to Benjamin Smedberg [:bsmedberg] from comment #8) > I was trying to verify that the fixup tool would actually do something > useful, and currently it doesn't appear to. I'm looking at making it work.
Attached file B2G minidump fixer (obsolete) —
Something changed in the minidump write in the past two years that makes the previous fixers not work in the current occurrence, even though the root problem is the same. So here is a new fixer program that should work better. I tested it on a minidump bsmedberg gave me, and it worked on that one. It should work on all of them. Build with g++ -I $m-c/toolkit/crashreporter/google-breakpad/src
OK think we're ready on the code side, going to file bugs to deploy this to stage+prod as part of Socorro 31 (release tomorrow)
Status: ASSIGNED → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
OK here is an example on stage that the batch job fixed, the original looks like this: https://crash-stats.mozilla.com/report/index/555cf5d3-372f-4ce3-83ff-2f0ae2121209 Now on stage after running the "fix broken dumps" job, it looks like: https://crash-stats.allizom.org/report/index/555cf5d3-372f-4ce3-83ff-2f0ae2121209
Flags: needinfo?(lars)
Fantastic, thanks!
Depends on: 822828
Depends on: 823171
QA verified on stage - Ted thanks for taking a look and rhelmer thanks for rocking this. The b2g reports look correct to me
Status: RESOLVED → VERIFIED
I re-ran my B2G crash list for yesterday - https://crash-analysis.mozilla.com/rkaiser/2012-12-18/2012-12-18.b2g.crashes.html - and looked at a few reports and they all look good now! Thanks so much for working on this, we finally have actionable data! \o/
Attached file B2G minidump fixer —
Ted found an issue with the previous iteration, that shows up on e.g. http://crash-stats.mozilla.com/report/index/2687199e-ff7f-4f1a-aa89-c17752121219 What happens is that the libraries are being fixed up properly, but other modules are altered in a way that dump_syms doesn't like. This new iteration fixes up these screwups, and avoids them from happening again.
Attachment #693570 - Attachment is obsolete: true
Commits pushed to master at https://github.com/mozilla/socorro https://github.com/mozilla/socorro/commit/465f5a04ea1035403e73c3843c693ab7a1edc62d bug 822432 - hbase API has changed a bit since this was last used (port stage branch patch to master) https://github.com/mozilla/socorro/commit/438d9f151c751b9170f023ddf6c27e7a7e9b6f4c Merge pull request #1024 from rhelmer/bug822432-hbase-api-change bug 822432 - hbase API has changed a bit since this was last used (port ...
Status: VERIFIED → REOPENED
Resolution: FIXED → ---
Depends on: 823595
Depends on: 823617
Anything left to do here?
Status: REOPENED → RESOLVED
Closed: 13 years ago → 13 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: