[Linux-cluster] CS4/ question about load on Heat Beat

[Patrick Caulfield]

Alain Moulle wrote:

> OK thanks Patrick.
> So to conclude on this point:
> the only way to avoid this problem in the case of
> two-nodes cluster is to have a second Ethernet
> network dedicated to the Heart-Beat of CS4
> (so that this network should never be over-loaded),
> isn't it ?
> Or do you have any other suggestion ?

That is what I would recommend.

> By the way, which is the value of timeout on Heart-beat
> check ?
> and how many times does it try to get the heart-beat
> in case of timeout ?

See /proc/cluster/config/cman/hello_timer &
/proc/cluster/config/cman/deadnode_timeout

The times are all in seconds and you can change them by writing values into
those files.

> Thanks again for your answers.
> 
> Alain Moullé
> 


-- 

patrick