[GE users] fatal error, run database recovery

Beadles, Jeff jeff_beadles at mentorg.com
Mon Mar 14 21:51:22 GMT 2005


    [ The following text is in the "iso-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

I ran a bunch of tests back when 6.0 came out, and the differences that we saw from bdb spooling and other changes (threading, I'm sure) were quite impressive.  With v5.3, we couldn't have more than about 12k jobs active at any given time, whereas with v6, 30k jobs and it didn't break a sweat.  I never did see what the top-end was.
 
I made the assumption that the bdb vs classic spooling was one of the differences that made things better.  We run what I think is a lot of jobs, 3 million+ per month, that are typically bursty.  (big flood of work, trickles down to nothing for a bit, and then another flood.)
 
Regards,  -Jeff

________________________________

From: Rayson Ho [mailto:raysonho at eseenet.com]
Sent: Mon 3/14/2005 1:34 PM
To: users at gridengine.sunsource.net
Subject: RE: [GE users] fatal error, run database recovery



The list is back!

I always use classic spooling for all my SGE6 installations. Is there a
real need for your site to use Berkeley DB spooling??

Rayson


> FYI, I ended up blasting everything, and reinstalling the grid
> master & clients.  I was able to recover part of the database,
> but there were several known things missing (like all.q), and
> who knows what else unknown.
>
>Regards, -Jeff

 




More information about the gridengine-users mailing list