Hi Serge,

> > But, I think that the log output which pgsql performs after ocf_run is 
> > noisy because ocf_run
> gives
> > error log in the case of pgsql.
> 
> I think we already addressed that nosiness in the monitor operation.
> As for start/stop I think it's rather critical to have an output
> information from pg_ctl in the log file.

The annoying log when monitor was finished normally was canceled by a q option.
However, ocf_run always outputs an error when an error happened.
And log is output in pgsql more in the back.

At the time of start/stop, it may be surely right that log after ocf_run 
appears.

If there is importance of the log output behind very much, I withdraw my patch.
And I explain that log increases in our user.

Best Regards,
Hideo Yamauchi.

--- Serge Dubrouski <[email protected]> wrote:

> Hi Hideo -
> 
> 2010/9/2  <[email protected]>:
> > Hi Serge,
> > Hi Lars,
> >
> >> > While I'm agree in general I'm not sure that it wouldn't be better to
> >> > modify ocf_run in a way that it would return actual error code form a
> >> > command that it ran instead of generic $OCF_ERR_GENERIC
> >>
> >> I'd be inclined to agree, but one needs to briefly review the resource
> >> agents to see whether this change would cause any ill effects.
> >
> >
> > It does not have the objection that ocf_run comes to give back a actual 
> > error code to the
> resource
> > agent.
> >
> > But, I think that the log output which pgsql performs after ocf_run is 
> > noisy because ocf_run
> gives
> > error log in the case of pgsql.
> 
> I think we already addressed that nosiness in the monitor operation.
> As for start/stop I think it's rather critical to have an output
> information from pg_ctl in the log file.
> >
> > ツ* There may be little influence on other resource agents.
> 
> There shouldn't be any impact. Agents usually just check the return
> code was zero in case of success or non zero in case of failure. pgsql
> though has that additional logic to check what was the real problem on
> the start of posgresql an expects to have a real exit code from
> pg_ctl.
> 
> >
> > Best Regards,
> > Hideo Yamauchi.
> >
> >
> > --- Lars Marowsky-Bree <[email protected]> wrote:
> >
> >> On 2010-09-01T20:26:09, Serge Dubrouski <[email protected]> wrote:
> >>
> >> > While I'm agree in general I'm not sure that it wouldn't be better to
> >> > modify ocf_run in a way that it would return actual error code form a
> >> > command that it ran instead of generic $OCF_ERR_GENERIC
> >>
> >> I'd be inclined to agree, but one needs to briefly review the resource
> >> agents to see whether this change would cause any ill effects.
> >>
> >>
> >> Regards,
> >> ツ ツ Lars
> >>
> >> --
> >> Architect Storage/HA, OPS Engineering, Novell, Inc.
> >> SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG N蟆・・nberg)
> >> "Experience is the name everyone gives to their mistakes." -- Oscar Wilde
> >>
> >> _______________________________________________________
> >> Linux-HA-Dev: [email protected]
> >> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> >> Home Page: http://linux-ha.org/
> >>
> >
> > _______________________________________________________
> > Linux-HA-Dev: [email protected]
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> > Home Page: http://linux-ha.org/
> >
> 
> 
> 
> -- 
> Serge Dubrouski.
> _______________________________________________________
> Linux-HA-Dev: [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
> 


_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Reply via email to