[Halld-offline] Final Lustre pool not recoverable [Re: CCPR 121020 UPDATE (Lustre/Infiniband issue of 12/31-1/1)]

[Mark Ito]

Bad news...see below.


On 01/19/2017 06:57 AM, ccpr_reply at jlab.org wrote:
> Here is an update to the help request you submitted.
> When you reply to THIS message, Please DO NOT include the original text below.
>
> Mod Date:   2017/01/19
> Mod Time:   06:57:23
> Mod User:   philpott
> Current State:   COMPLETE
> STATE changed from (WORKING) to COMPLETE.
>
>
> --------------------------------------------------------------
> News Jan. 18
> Lustre/ZFS recovery complete; final pool not recoverable
>
> After Intel support was able to recover 4 of the 5 problematic ZFS
> storage targets that became unavailable in Lustre Dec 31, the last
> target is corrupt and cannot be recovered, holding a total of ~8 TB of
> data. Users can remove problematic files on this target from Lustre
> with the "unlink" command.
>
> 									
> --------------------------------------------------------------		
> Here is a copy of your Original Request:
> 	
> Email:     marki at jlab.org
> Name:      Mark Ito
> Username:  marki
> Staff:	   philpott
> Platform:  other,Netscape,20100101 Firefox
> Building:  12_1
> Room:      A104
> Hostname:  129.57.73.156
> Category:  SCIENTIFIC COMPUTING
> Subject:   Lustre/Infiniband issue of 12/31-1/1
> Submitted: 1/3/2017 8:50 AM
> 			
> Request:
> Is there an update on this problem? Should we expect normal operation?
> I am seeing problems on /work/halld on ifarm1401. 'ls' not responding
> from /work/halld/pull_request_test .
>
>
> The following link will bring you to a list of open CCPR's
>
> http://mis.jlab.org/mis/ccpr/ccpr_user/ccprframe_user.html
> 	

-- 
marki at jlab.org, (757)269-5295