521231 - update Breakpad code on production Socorro processors to SVN r409

Status: VERIFIED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[(not currently active) Ted Mielczarek]

I landed Jeff's patch to get us better crash stacks from minidump_stackwalk in some cases in upstream Breakpad SVN, r409. I'd like to get the copy of minidump_stackwalk used on the production Socorro processor updated to one built from this revision so we can start getting better stacks for some of our crashes.

Comment 1

[Aravind Gottipati [:aravind]]

Will push this to staging so we can test it there first.

Comment 2

[Aravind Gottipati [:aravind]]

6aceb4e5-68aa-4007-80a0-9a6412091009
da936478-22f9-4f65-a543-f481d2091009
7a8851b7-b27d-413e-aa25-a0cba2091009
e2f896ea-dfbb-4d72-83c8-b40fb2091009
8a0c0d96-ebcd-48f0-be93-4520d2091009
7d7e814f-c7c5-4097-8260-a39502091009
37066b54-5f71-42e1-af9b-ceb0c2091009
2fd509d8-48b5-468e-80bb-a1b8b2091009
5c5edba4-d766-4b0a-b064-4aecb2091009
8a05a9b5-9dfe-42e9-811d-e72a12091009
9ea248b2-f1b3-416b-9b64-8de952091009
9158e330-b5a7-4441-95b1-681152091009

Processed the above crashes in staging with the new stackwalk.  Once you guys okay this, I can push to prod.

Comment 3

[Jeff Muizelaar [:jrmuizel]]

The first one (6aceb4e5-68aa-4007-80a0-9a6412091009) doesn't seem any better. In fact none of them seem much better than what I'd expect before the patch.

Comment 4

[Jeff Muizelaar [:jrmuizel]]

aravind: It would be great if we could get a hold of the minidump for 6aceb4e5-68aa-4007-80a0-9a6412091009 so we can figure out why that one isn't producing better results.

Comment 5

[Aravind Gottipati [:aravind]]

Sorry, the dump numbers I have you were autogenerated by the system and were deleted as soon as they were processed.  However, I am now storing the original file to the processed name mappings.  Please ping me when you are ready to test again and I will get you the real name-> generated name mappings so it will be easier to track the original files.

Comment 6

[(not currently active) Ted Mielczarek]

Can we try this again?

Comment 7

[(not currently active) Ted Mielczarek]

I reconvened with aravind and we had a miscommunication. I had told him earlier about a Breakpad SVN branch I had done some other work on, and he thought that was the code we wanted to use. He's updated to SVN trunk and is re-running this. My manual results look good, I found a 0x0 crash in that list and processed it with the new code and got a more useful stack.

Comment 8

[(not currently active) Ted Mielczarek]

This looks good. For example:
http://crash-stats.stage.mozilla.com/report/index/bp-d5990bbc-a4d5-4188-94f5-d22e82091013
http://crash-stats.stage.mozilla.com/report/index/bp-b17cedeb-c948-43fd-be6a-7f8762091013
http://crash-stats.stage.mozilla.com/report/index/bp-6cb1098d-bfb2-467e-99ce-e352c2091013
http://crash-stats.stage.mozilla.com/report/index/bp-bf9b9c57-8f13-4d8d-8ab4-21eb02091013

We should put this into production.

Comment 9

[Aravind Gottipati [:aravind]]

Running in production now.

Comment 10

[(not currently active) Ted Mielczarek]

Looks good, there are no crashes with @0x0 signatures in the past 12 hours!