skip repl record (old)

["Arredondo, Tomas" <tArredondo@unispherenetworks.com>]

Hi all,

I am getting this replication error even though I've restarted my master and
slave slapd as well as slurpd.  Any ideas on how to reset this timestamp
stuff?  I looked at the code and it was a bit criptik.  This is in 1.2.11 (a
legacy system). 

Replica localhost:33334, skip repl record for
SERVICENAME=VOIP,RETAILERNAME=9999,O=USERS,O=UMC (old)
Replica localhost:33334, skip repl record for
SERVICENAME=VOIP,RETAILERNAME=9999,O=USERS,O=UMC (old)

Thanks,
Tomas

REF:

/*
 * Return 1 if the Re's timestamp/seq combination are greater than the
 * timestamp and seq in the Ri's ri_stel member.  In other words, if we
 * find replication entries in the log which we've already processed,
 * don't process them.  If the re is "old," return 0.
 * No check for NULL pointers is done.
 */
static int
isnew(
    Ri  *ri,
    Re  *re
)
{
    int x;
    int ret;

    /* Lock the St struct to avoid a race */
    sglob->st->st_lock( sglob->st );
    x = strcmp( re->re_timestamp, ri->ri_stel->last );
    if ( x > 0 ) {
        /* re timestamp is newer */
        ret = 1;
    } else if ( x < 0 ) {
        ret = 0;
    } else {
        /* timestamps were equal */
        if ( re->re_seq > ri->ri_stel->seq ) {
            /* re seq is newer */
            ret = 1;
        } else {
            ret = 0;
        }
    }
    sglob->st->st_unlock( sglob->st );
    return ret;
}

Tomás V. Arredondo
Member of Technical Staff
Unisphere Networks
7800 Congress Ave., Suite 100, Boca Raton, Fl 33487
T 561.981.7051 F 561.981.7001
www.unispherenetworks.com