Ed Morley [:emorley] Reporter
	Description • 12 years ago

Bug 768156, bug 767906 & bug 768483 lack proper stacks & are all from the same slave (talos-r3-w7-038).

	Ed Morley [:emorley] Reporter
	Comment 1 • 12 years ago

It would seem that slave isn't very well.

Of the 38 new [orange] crashes filed in the last 7 days:
https://bugzilla.mozilla.org/buglist.cgi?keywords=crash%2C%20;keywords_type=allwords;list_id=3531381;status_whiteboard_type=allwordssubstr;chfieldto=Now;query_format=advanced;chfield=[Bug%20creation];chfieldfrom=2012-06-19;status_whiteboard=[orange]

20 of them originated from this slave:
https://bugzilla.mozilla.org/buglist.cgi?keywords=crash%2C%20;keywords_type=allwords;list_id=3531378;status_whiteboard_type=allwordssubstr;chfieldto=Now;chfield=[Bug%20creation];query_format=advanced;chfieldfrom=2012-06-19;status_whiteboard=[orange];longdesc=talos-r3-w7-038;longdesc_type=allwordssubstr

Bill, seems like some of those JS crashes may not be real after all :-(

	Ed Morley [:emorley] Reporter
	Comment 2 • 12 years ago

Not yet filed:
https://tbpl.mozilla.org/php/getParsedLog.php?id=12980564&tree=Mozilla-Inbound

	Ed Morley [:emorley] Reporter
	Comment 3 • 12 years ago

https://tbpl.mozilla.org/php/getParsedLog.php?id=12982610&tree=Mozilla-Inbound

	Ed Morley [:emorley] Reporter
	Comment 4 • 12 years ago

(More unfiled):
https://tbpl.mozilla.org/php/getParsedLog.php?id=12929167&tree=Fx-Team
https://tbpl.mozilla.org/php/getParsedLog.php?id=12952071&tree=Fx-Team
https://tbpl.mozilla.org/php/getParsedLog.php?id=12973085&tree=Fx-Team
https://tbpl.mozilla.org/php/getParsedLog.php?id=12971305&tree=Fx-Team
https://tbpl.mozilla.org/php/getParsedLog.php?id=12981379&tree=Fx-Team

	Ed Morley [:emorley] Reporter
	Comment 5 • 12 years ago

Please may we remove this slave from production.

	Bill McCloskey [inactive unless it's an emergency] (:billm) Assignee
	Comment 6 • 12 years ago

Actually, we see users who have crashes like and we can never solve them. Could we try to figure out what's wrong with this slave so we can help our users?

	Ed Morley [:emorley] Reporter
	Comment 7 • 12 years ago

Armen, please may you give billm access to the slave after taking it out of production. Thanks! :-)

	Armen [:armenzg]
	Comment 8 • 12 years ago

I disabled the slave on slavealloc and added a note.

billm, would you be able to look what is wrong with this slave?

	Bill McCloskey [inactive unless it's an emergency] (:billm) Assignee
	Comment 9 • 12 years ago

(In reply to Armen Zambrano G. [:armenzg] - Release Engineer from comment #8)
> I disabled the slave on slavealloc and added a note.
> 
> billm, would you be able to look what is wrong with this slave?

Yes, I'd like to try. At the very least, it would be interesting to run a memory test. It may just be a hardware malfunction. Can you please send me a login?

	Armen [:armenzg]
	Comment 10 • 12 years ago

I asked IT to get you access.
You should have the machine ready tomorrow morning.

	Armen [:armenzg]
	Comment 11 • 12 years ago

I have sent the credentials to Bill.

	Mike Taylor [:bear]
	Comment 12 • 12 years ago

removing this from Critical to Major as it is in dev's hands to work with

	Chris Cooper [:coop] (he/him)
	Comment 13 • 12 years ago

Bill: any update on the status of this slave? Found anything? Are you done with it?

	Bill McCloskey [inactive unless it's an emergency] (:billm) Assignee
	Comment 14 • 12 years ago

I ran a memory test and didn't find anything. I tried running a few tests and none of them failed.

However, I don't think we can add this slave back to the pool. It will likely just cause more failures.

	Chris Cooper [:coop] (he/him)
	Comment 15 • 12 years ago

(In reply to Bill McCloskey (:billm) from comment #14)
> I ran a memory test and didn't find anything. I tried running a few tests
> and none of them failed.
> 
> However, I don't think we can add this slave back to the pool. It will
> likely just cause more failures.

OK, thanks for trying. I'll get IT to re-educate the machine in bug 776924.

	Chris Cooper [:coop] (he/him)
	Comment 16 • 12 years ago

This slave has been fixed, at least in theory. If this recurs, please re-open and we'll just decommission this slave.