On 4/24/06, Moe <[EMAIL PROTECTED]> wrote: > The entries in the slave ldap start getting lost then after a while ldap on > the slave crashes. When this happens i restore the entries from master, > ldap slave gets replicated fine for couple of months, then suddenly start > losing entries. Master reject logs don't report anything then ldap slave > crashes again. > > Moe > > Moe <[EMAIL PROTECTED]> wrote: > Hi Everyone, > > I have master slave replication using slurpd. I noticed that after a while > the replica site is getting out of sync with the master site. The reject log > on the master ldap does not have any data corresponding to rejected entries. > Im suspecting some king of network > issue.The entries were in sync for like 3 months and i noticed recently > that the slave is missing like 20 entries. Any suggestions? > > Thanks > Moe
My guess is that slapd is probably crashing without writing the entries it had in memory to disk. There are some back-bdb tunables for keeping the disk in sync. You might be able to gcore slapd when it gets out of sync/ready to crash and see if something is hung, or whatever.
