if you see heartbeat timeouts on disk boot with scheduler=deadline
please and see if that solves your problem
we have serious issues with cfq
Alexei_Roudnev wrote:
> I updated scripts and manual for Oracle 10.2 @ OCFS && ASM @ SLES9 Sp3
> (http://ftp.portera.com/Linux/);
> important change rely to OCFSv2 timeouts (taken from
> here -http://www.oracle.com/technology/pub/articles/hunter_rac10gr2_2.html#1
> 6 ). Problem is that OCFSv2 detects failure and panic if it cannot make IO
> on the voting area of disk in 12 seconds, which is smaller than many IO
> timeouts around (NetApp takes about 10 - 20 seconds to takeover, for
> example).
>
> Btw, I wonder, why SuSe don't change defaults for panic mode on servers - it
> is obvious that servers should have defaults as
>
>  kernel.panic=1
>  kernel.panic_on_oops=1
>
> (who need server which do not restart in case of panic?)
>
> Timeout issue can explain few oops seen by OCFSv2 users (together with panic
> settings).
>
> PS. After all, it is all pretty stable on iSCSI. FC and Multipath is another
> issue - require careful configuration (I was not able to configure it for
> netApp FSP and old Qlogics 1Gbit HBA, and system freeze easily with and
> without different 'oops').
>
>
>  


--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting