Today at 1:18pm, Bill Dossett wrote:
Hi,
Hello,
Not sure what I've done wrong here, if this sounds
familiar to anyone? I've pretty much messed things
up pretty badly by now I guess... but anyway,
Ok, from your description, it sounds like you have completely destroyed
the slave's database and need to start from ground zero.
At this point, the steps I would take are:
1) stop slapd on the slave (if it's running)
2) stop slapd and slurpd on the master
3) use slapcat on the master to dump the database
4) make sure the replica log files on the master are empty
5) start slapd/slurpd on the master
6) transfer the slapcat generated file (step 3) to the slave
7) use slapadd to load the database on the slave
8) start slapd on the slave
9) change an entry on the master and see if it shows up on the slave
If step 9 fails, something is messed up in your replication process,
check the replication logs for errors
You asked where does the master keep all the information it needs to
replicate. It doesn't. It hands it all to slurpd (via the replication
log file) at the time the change is made and promptly forgets about it.
What slurpd does with the information if the update fails (because the
slave isn't there)... I've never tested, someone else will have to
answer that.