slurpd queries and Failover

["Adrian Black" <adrian.black@itouch.com.au>]

1.What dictates the check period for slurpd.
2.How does it keep track of where it is in a replog file? Is it possible at
any time to find out how far it is in a replog file and thus know what it
has to do.
3.Does the replog file just keep growing and growing and is only be
truncated by human intervention.
4. (related to 2) when run in one shot mode does it try to update all the
changes in the supplied file or does it know how far it got at some other
time with the target Slave.

Why do I need to know this? I need to implement a failover for the LDAP
server which we have that acts as an address book for an application that we
have.
The current plan is as follows...
Normal Operation
1 Production and Failover are 2 identical machines. Failover kicks in when
Production dies.
1.1 Production LDAP is Master
1.2 Failover LDAP is Slave and recieves updates from the Master, thus at
failover the slave contains all
the information that the Master has (or almost all since it is possible that
failover can occur in between
an update on the master without the replication server having updated the
Slave).

2.Production goes down and failover occurs.

3 Failover LDAP is stopped and restarted as a normal LDAP server neither
Slave nor Master.
3.1 Updates occur on the Failover LDAP as normal.

4 Production comes back up

5 Service Unavailable Screen is shown
5.1 Failover LDAP is stopped and its databasefiles are copied to replace the
ones of the Master which is also not currently running.
5.2 Master is started
5.3 Web site goes live
5.4 Failover is started as Slave again
5.5 Replication server is restarted.


Problems:
1 Updates not replicated to the slave before failover are lost.
2 Service Unavailable screen
3 Most probably will have to be done by human control and observation i.e.
would be hard to implement automatically

Thanks