[Ocfs2-users] About heartbeat threshold and drbd ...

Sébastien CRAMATTE s.cramatte at wanadoo.fr
Mon Nov 6 02:56:00 PST 2006


Hi

I've setuped  drbd 0.8pre6 over OCFS2 and I've got some stability
problems ...
I've submited somes erros as bug on  ocfs2 bugzilla but doesn't appear I
supose that are already referenced ....

I've send this to the drbd ml  and  "Lars" talk about  

ocfs2 self-fencing threshold of 7 ... how can I change it to 10 ? 
Sorry for this basic question but I'm newbie in OCFS2


Regards

Lars Ellenberg a écrit :
> / 2006-11-05 21:37:52 +0100
> \ Sébastien CRAMATTE:
>   
>> I've speak too quickly ...
>> 5 mn after ...
>> I've copied  (using rsync)  my /home directory to  the mounted  ocfs2  
>> folder ... works quite well
>> But when I've tried to delete somes files inside the cluster  .... 
>> Kernel Panic !!!
>>     
>
> well, if it was ocfs2 self-fencing,
> make sure that your timeout (ocfs heartbeat retry) settings
> are suitable for your setup... the default ocfs2 self-fencing threshold
> of 7 ocfs2 heartbeat retries (effective kernel panic after 12 seconds) 
> is _way_ too short for a default drbd timeout of 10 seconds.
>
>   
>> drbd0: Split-Brain detected, dropping connection!
>>     
>
> you have to decide which node has to be sync target,
> make it secondary, and then tell it to connect
> and "discard" its data:
>
> # drbdadm -- --discard-my-data connect whatever
>
>   



More information about the Ocfs2-users mailing list