Hello!

I have played with replication and failovers and it worked very pretty,
but I have noticed a strange thing that shouldn't have happened.

It was inconsistent user database on the slave server when I shut
down the scm and tested authentication.
Files *.db had rather old (and matching with the master) timestamp but md5
were not the same as on the master.
[In pdbtool on the slave I could do list and it was right but search for
some names ('i' command) was impossible. There were just two usernames in
the database and search worked for one of them, not for the other.]

When I tried to troubleshoot, after the first change on the scm the files
got in sync and the problem disappeared :-/

I suspect there may be race conditions while copying the db files
across the servers.

Of course I did not mess with the user database in the replica, neither
had I had dirty shutdowns.

Anyway, my problem has gone now. But im logging on the list so that
the next time somebody sees that thing, we have a reference.

Best regards,
--
Ivan