Hi, We recently encountered the same kind of issue. It turns out the PostgreSQL library (libpq) was not compiled with --enable-thread-safety, and this causes random breakage in slon. The slony docs say you can't compile in this situation, but the configure script checks this only on some platforms (AIX, Solaris), see line 6823.
Looks like the check should be performed in more cases, probably all? Not sure this is the issue that is affecting you, but it's worth checking (check pg_config for any reference to threads to find out). Jacques. At 12:24 12/06/2007, James Tucker wrote: >Hi Folks, > >I've recently configured slony to replicate our production >databases. One of the databases has ceased replicating and I can >see events queueing in sl_status. On the subscriber node I have one >connection "idle in transaction" to postgresql running as the slony >user. I see no errors, warnings or even notices in the logs, just >debug level sync events which appear to be going through ok. > >I've tried restarting the slon processes however the idle in >transaction connection just keeps reappearing, I am unable to >restart the postgresql service on either node because of availibity >issues. Any hints on where to go from here would be very welcome. > >Cheers >James >_______________________________________________ >Slony1-general mailing list >[email protected] >http://gborg.postgresql.org/mailman/listinfo/slony1-general _______________________________________________ Slony1-general mailing list [email protected] http://gborg.postgresql.org/mailman/listinfo/slony1-general
