Bug#638316: AW: Same problem here, running squeeze on VMware vSphere 5 ESXi
Hi Jonathan,
unfortunately i dont know a reliable way to reproduce this. About a week ago
we did run some performance- and loadtests which caused very high load over a
period of more than 48 hours and nothing happened. This week we encountered
the described problem on the database system, but only temporarily.
One day before we had a crash of the application server, the stack traces
there seemed to look quite similar. As i'm not a specialist in kernel
development i thought it might be the same bug, but now i realized that there
isnt any hint for blocking tasks. Instead there are lines like "BUG:
scheduling while atomic: sshd/22743/0x10000001". So i think you are right and
i should file one or maybe two separate bugs.
Best regards
Sebastian
-----Ursprüngliche Nachricht-----
Von: Jonathan Nieder [mailto:jrnieder@gmail.com]
Gesendet: Fr 30.03.2012 00:45
An: Sebastian Schäfer
Cc: 638316@bugs.debian.org
Betreff: Re: Same problem here, running squeeze on VMware vSphere 5 ESXi
Hi Sebastian,
Sebastian Schäfer wrote:
> The system is running a postgresql9 Server (Version: 9.0.6-1~bpo60+1) with
> two hotstandby Clients, using streaming replication.
Since you are not using Xen, this is most likely not the same bug.
Do the processes remain in the D state, or are they only temporarily
so in some time of particularly high load? If you want to pursue this
(especially if you know of some reliable way to get processes stuck in
the D state), please file a separate bug.
Thanks for writing,
Jonathan
Reply to: