937313 - please run hardware diagnostics on and reimage talos-r3-fed-057

Status: RESOLVED FIXED

(Infrastructure & Operations :: DCOps, task)

Description

[[e-mail] (:bhearsum)]

It's having read-only filesystem issues.

Comment 1

[John Pech [:jpech-lbc]]

Keeps on crashing and restarting while running Diagnostics. Will resume on fix tomorrow.

Comment 2

[John Pech [:jpech-lbc]]

Machine still crashing during diagnostics test when checking for memory phase. Will start the re-image process to see if will go through.

Comment 3

[Vinh Hua [:vinh]]

Diagnostics indicates the following error code:
4M0T/1/40000002: Master -0

I was not able to search online for the exact code details. Apparently Apple does not provide this to the general public.  *shame on them*
  
I'm thinking it might be a motherboard issue.  Let me know how you want to proceed.

Comment 4

[John Hopkins (:jhopkins)]

vinh: can you try the PRAM and SMC reset procedures? (https://discussions.apple.com/thread/1404579?start=0&tstart=0) and retry the diagnostics?

Comment 5

[Vinh Hua [:vinh]]

Sure I'll give it a shot.

Comment 6

[Vinh Hua [:vinh]]

PRAM and SMC reset did not resolve the error:  4M0T/1/40000002: Master -0 issue.

Comment 7

[Vinh Hua [:vinh]]

We've just decomm the talos-r3-xp/w7 minis.  I can reimage one of those to replace this faulty one.  Or do you prefer to decomm this host?

Comment 8

[John Hopkins (:jhopkins)]

Armen: should we replace this failed talos-r3-fed slave with one of the talos-r3-xp/w7 minis or just decomm without replacement?

Comment 9

[Armen [:armenzg]]

vinh, it would be great if we could re-purpose a bunch of Win7/WinXp machines as Fed64/Fed machines.

Let me include you on an email thread.

Comment 10

[Vinh Hua [:vinh]]

Armen,
Haven't heard any further discussion on the re-purpose of win7/xp to Fedora machines from the email.  In the meantime should r3-fed-057 be decomm?

Comment 11

[Armen [:armenzg]]

Hi vinh,
I'm encouraging developers to move us to AWS before the end of Q1.
Even if we imaged one of those win7/winxp machines as fedora machines we don't know:
1) if the image still works as expected (I don't know when was the last time we re-imaged a fed/fed64 machine
2) we don't have puppet anymore for these machines

I expect that we will leave the win7/xp machines untouched until we're closer to the end of Q1 and we see how running the jobs on AWS is going.

If we can't fix fed-057 easily (I don't know what comment 6 means) let's decommission it.

Thanks for your patience and keeping me accountable.

Comment 12

[Vinh Hua [:vinh]]

:arr - Can you remove this host from nagios for decomm?

Comment 13

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

removed from nagios

Comment 14

[Vinh Hua [:vinh]]

Host has been decomm.