AW: slurpd doesn't recognize that the slave is down

["Harms, Hendrik (TOJAQ)" <extern.hendrik.harms@volkswagen.de>]

Hi,

You're right. The replication works fine!  :-)

I made some more tests with shouting down the slave or/and the slurpd.
Then both are up again the changes are replicated. I was just confused,
cause I can't find a file where the slurpd writes the changes into when a 
slave is down. 

I am still curious:
Where does the slurpd persist pending replications?
They are deleted from the replogfile but not written into the slurpd.replog.

Thanks, Hendrik

On Wed, 19 Feb 2003, Held, Oren wrote:
>
> Hi,
> 
> I think that you're a bit confused. The *.rej files contain ldap
> commands (ldifs actually) which were REJECTED by one of the slaves (i.e.,
> permission denied).
>
> If a slave is down, the replication log just gets bigger and bigger, and
> when the slave would wake up, slurpd should send that big replication log
> to him. Replication log should be emptied all the time in healthy systems.
> 
> That is, if I understood your question well.
> 
> - Oren
>
> On Wed, 19 Feb 2003, Harms, Hendrik (TOJAQ) wrote:
>
> > I'm using openldap-2.0.27 in a master slave configuration on solaris
> > The replication works. But when the slave ldap is down, the slurpd
> > did not write anything into the *.rej files. All changes are written
into
> > the slurpd.replog file like those the slave is up.
> > How could I find out what changes are sent during the downtime of
> > the slave?
> >
> > Thanks,
> > Hendrik
> >