[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: What slurpd is waiting for?



hi alls:

ok, im using the 2.0.19 version. for me, if i do changes at master the
replica is do too (seems like the slurpd dont wait to full a buffer).

do you have standard slapd.conf? or did you make changes at it?


kind regards

----- Original Message -----
From: "Al Nikolov" <al@iac.spb.ru>
To: "Agustin" <torrest@andaluciajunta.es>; <openldap-software@OpenLDAP.org>
Sent: Monday, November 04, 2002 11:40 AM
Subject: Re: What slurpd is waiting for?


> Hello
>
> Yes, i think so. My status file is empty because the master is never
> connected to the slave. Sure it is! My slapd only had start to producing
> replicas and my slurpd only had start to 'gathering' it to the private
> replog. When it'll be fat enough it'll connect to the slave to flush the
> replicas. We could wait for years before it'll just try to unload!
>
> On Monday 04 November 2002 11:44, Agustin wrote:
> > hi alls!
> >
> >  about the number of changes that are necessary to run the slurpd i dont
> > have any idea. but i  think that your problem is other and i think this
> > because your status file is empty.
> >
> > the status file must have one string for replica defined at slapd.conf.
the
> > format of this string is the next:
> >
> >     host of slave : port of slave : time of last change aplicated to
this
> > slave : one code that i dont know
> >
> > if your slurpd.status is empty i think that you dont connect the master
to
> > slave never.
> >
> > agustin.
> >
> > PD: anypone have any idea about the last code of the string that i have
> > described before?
> >
> > ----- Original Message -----
> > From: "Al Nikolov" <al@iac.spb.ru>
> > To: <openldap-software@OpenLDAP.org>
> > Sent: Saturday, November 02, 2002 4:31 PM
> > Subject: What slurpd is waiting for?
> >
> > > Hi
> > >
> > > I'm using openldap 2.0 and just set up a simple replication between 1
> >
> > master
> >
> > > and 1 slave. I got following picture: when i modify master slapd it
> > > writes replog, then slurpd watched it, truncates this replog and write
> > > its own replog in the 'temporary directory'. It's all.
> > >
> > > slurpd.replog now contains the replica
> > > slurpd.status is empty
> > > slave doesn't receive the replica never
> > >
> > > man slurpd(8) says: "Normally, slurpd processes the replog file and
then
> > > watches for more replication entries to be appended." Not too
> > > informative. How many replicas to be appended before slurp will go to
> > > bind to the
> >
> > slave?
> >
> > > AND what should i do to periodically 'flush' saved replicas - without
> >
> > waiting
> >
> > > for 'fulfilling' of the slurp 'buffer'?
> > >
> > > Thanks
>