[Clusterusers] keg rebuild attempt

Wm. Josiah Erikson wjerikson at hampshire.edu
Thu Mar 5 14:46:59 EST 2015


I have replaced the failed disks, recreated the RAID, installed Ubuntu
14.04, run updates, and am now copying the backed-up /helga over to the
new keg and waiting for the RAID array to resync. These things will
probably take all night. Seeing as how I'm sick and got in at like 2AM
last night, I'm going home and will pick this up tomorrow - I'll keep
you updated.
    -Josiah


On 3/5/15 11:51 AM, Wm. Josiah Erikson wrote:
> Hi folks,
>     I have failed in my attempts to get keg's RAID array back. I will be
> rebuilding keg with Ubuntu 14.04LTS, and then manually configuring
> everything (which shouldn't be too hard) and copying things back from
> backups. I hope *fingers crossed* to have everything back up by the end
> of tomorrow. If I don't, I will work on it over the weekend.
>     Until then, tractor, helga, and the render farm in general will
> continue to be down. My sincerest apologies for this inconvenience!!
>     -Josiah
>
>
> On 3/5/15 10:02 AM, Wm. Josiah Erikson wrote:
>> Hi guys,
>>     I am going to start working on getting keg back online. This will
>> happen in two stages:
>>
>>     1. I am going to attempt to get the RAID array to re-assemble, so we
>> can start again right where we left off. My initial attempts don't look
>> promising, but I'm going to keep trying
>>     2. Failing that, I am going to have to rebuild from scratch and then
>> restore the backed-up data
>>
>>     IMPORTANT INFO:
>>
>>     We do not back up everything. The svn repo is of course backed up,
>> and all of /helga/tube is backed up, but only because we made a special
>> exception. Unfortunately, the only thing in /helga/<yourfolder> that is
>> backed up is anything that you put in /helga/<yourfolder>/prot, which
>> means that any existing renders, etc, will be gone if I have to go with
>> #2, which looks increasingly likely.
>>
>>     The idea here is that only machine time is required to rebuild what
>> you've lost if you've put all the source material in the svn repo. I
>> hope that this is the case, but it still kinda sucks.
>>
>>     I forgot my phone today, but feel free to stop by the server room or
>> shoot me an email back.
>>

-- 
Wm. Josiah Erikson
Assistant Director of IT, Infrastructure Group
System Administrator, School of CS
Hampshire College
Amherst, MA 01002
(413) 559-6091



More information about the Clusterusers mailing list