829014 - repair talos-r4-snow-041

Status: RESOLVED FIXED

(Infrastructure & Operations :: DCOps, task)

Description

[[e-mail] (:bhearsum)]

Hardware diagnostics came up clean on this slave, but even after a re-image it's still failing tons of tests. Can we do any repairs even when diagnostics come up clean?

Comment 1

[Van Le [:van]]

>Hardware diagnostics came up clean on this slave, but even after a re-image it's still failing tons of tests. Can we do any repairs even when diagnostics come up clean?

:bhearsum, I'm not sure what you mean. Do we spend hundreds of dollars randomly replacing parts? You have to keep in mind we're short staffed as everyone else so we don't really have time to play around with these boxes. Do you have any logs that would help us out?

BTW, what is the threshold for decommissioning these nodes?

Comment 2

[Amy Rich [:arr] [:arich]]

Should we just collect all of the ones that have come up with clean diags but that seem to be burnign builds and send them off to Computer Care for further diagnosis?  I'm not sure if they'd do more than we're doing in house to diagnose...

Comment 3

[Vinh Hua [:vinh]]

Diagnostic test resulted with:

Completed:Volume:Jan 24, 2013 1:40 PMMacintosh HD (319.73 GB /)
CHECKED
3 suspected files found out of 1557 total files checked.

Suspected files cannot be repaired and should either be deleted or replaced with a known good backup copy.
File is corrupt or is an unsupported file format:
/Users/cltbld/Library/Caches/TemporaryItems/device-storage-testing/RBERRSTV0CBKNZQXF45H.png

Unknown character '{' (0x7b) in <integer> on line 10:
/Library/Ruby/Gems/1.8/gems/facter-1.5.6/conf/osx/PackageInfo.plist

Unknown character '{' (0x7b) in <integer> on line 10:
/Library/Ruby/Gems/1.8/gems/puppet-0.24.8/conf/osx/PackageInfo.plist

The File Structures test checks a wide variety of file types to ensure that they are structured correctly. If a file is not, the program reports the name of the file.



Looks like some file structures are corrupt.  I will reimage the host tomorrow.

Comment 4

[Amy Rich [:arr] [:arich]]

These can be reimaged remotely as long as you can get root access to the machine.  I've kicked off a reimage.