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

Re: slapd 1.2.13 stops replicating



It's my production master, so I don't have much logging turned on.
Backend is BDB 2.7.7.  Server certainly acts like it thinks it gets a
MON operation, since it changes the data in ther master DB.  But nothing
appears in the replog.

Is there any chance this could be thread related?  While the FAQ-O-Matic
describes (and I use) the Solaris threaded build hints, I note a mention
that Solaris threads are preemptive and unsuitable for OL 1.2.x.  Wonder
if this is an artifact of that...  but I reach.

-Alan

wes@umich.edu wrote:
> 
> Ok.  Are you getting a log line that says that the MOD has been requested?
> What backend are you using?  The replog() function pretty much gets called
> directly after the backend's modify function, e.g.
> 
>     if ( (*be->be_modify)( be, conn, op, ndn, mods ) == 0 ) {
>         replog( be, LDAP_REQ_MODIFY, ndn, mods, 0 );
>     }
> 
> So, if the backend indicates it, slapd write the the modification to the
> replog.  What kind of logging do you have turned on?
> 
> :wes
> 
> On 2/20/02 9:45 PM -0700 asparks@quris.com wrote:
> 
> > Well, I've rebuilt the database many times... still chronically fails.
> > Removed slurpd.status before, also.  I don;t think the problem is in
> > slurpd anyway -- shutting slurpd off and watching the replogs show no
> > change logs getting written.  Until I restart slapd, that is.
> > -Alan
> >
> > wes@umich.edu wrote:
> >>
> >> Just a guess, but when I get systemic problems like this, I generally
> >> have to rebuild the DB.  Is it possible that your slurpd.status is
> >> getting corrupted?
> >>
> >> :wes
> >>
> >> On 2/20/02 6:31 PM -0700 asparks@quris.com wrote:
> >>
> >> > Still running 1.2.13, cuz of feature probs going to 2.0.x.  Beside the
> >> > point.
> >> >
> >> > I've been having a major problem with my master server on Solaris 7
> >> > just stopping replication.  It appears that the slapd accepts
> >> > modifications, but just doesn't write them to the replog.  A restart
> >> > of the server clears the problem for a while, until it recurs in a day
> >> > (or less).
> >> >
> >> > Has anyone had any experience with this happening, and found a
> >> > solution?
> >> >
> >> > Compiled with GCC 2.95.2, with the following options:
> >> > CPPFLAGS="-I/usr/local/include" \
> >> > LDFLAGS="-L/usr/local/lib" \
> >> > LIBS="-lgnuregex -lpthread -lposix4" \
> >> > ./configure \
> >> >         --prefix=/opt/ldap \
> >> >         --with-threads \
> >> >         --enable-shared \
> >> >         --enable-phonetic \
> >> >         --enable-shell
> >> >
> >> > Thanks in advance for any advice.
> >> > -Alan
> >> >
> >> > --
> >> > Alan Sparks, Sr. UNIX Administrator     asparks@quris.com
> >> > Quris, Inc.                             (720) 836-2058
> >
> > --
> > Alan Sparks, Sr. UNIX Administrator     asparks@quris.com
> > Quris, Inc.                             (720) 836-2058

-- 
Alan Sparks, Sr. UNIX Administrator	asparks@quris.com
Quris, Inc.				(720) 836-2058