[Linux-HA] Removing "hbgenmethod time" option

Ragnar Kjørstad linux-ha at ragnark.vestdata.no
Wed Mar 14 14:32:53 MDT 2007


On Wed, Mar 14, 2007 at 01:27:58PM -0600, Alan Robertson wrote:
> > Why don't we change heartbeat to use hb_generation if it exists, and
> > start the hb_generation counter on time_t if it doesn't?
>
> ...
> 
> That's mostly reasonable.  I _like_ starting it at one aesthetically,
> (it gives me a feeling for how many reboots I've done) but agree with
> most of your reasoning.

Just to clarify; you agree with the proposal _inspite_ of liking it to
start on 1, right?

Yes, I too agree the small and nice numbers :)

> However, I wouldn't remove the directive, since failure to write the
> hb_generation file should be logged as an error.  And, for the diskless
> systems that "error" is expected - therefore not really an error.

So, with "hbgenmethod file" failure to write the hb_generation file
would be a fatal error, with "hbgenmethod time" it would not even be
attempted?

And "hbgenmethod file" would still be the default?

> Or, if the errno is EROFS then I suppose I could ignore it, or at least
> downgrade it to an informational message.

Yes, something like this would be needed for diskless to work out of the
box.

> Could you make a bugzilla for this and assign it to me?

Done. #1522.

-- 
Ragnar Kjørstad
Software Engineer
Scali - http://www.scali.com
Scaling the Linux Datacenter


More information about the Linux-HA mailing list