I appreciate that this is a few months old however we had apparently
identical symptoms but on a Windows platform.
In the end we discovered that the reason the WAL archives that existed on
the slave were not being restored was because the Windows account postgres
was running as (Network Service) did not have permissions to access the WAL
Archive directory. Once we gave it permissions we started seeing restoration
messages in the log:
"2013-09-12 13:32:42 BST LOG: restored log file "0000000100000090000000FD"
from archive"
Additionally, when we gave Network Service write permissions to the WAL
Archive directory, our pg_archivecleanup command also started succeeding
after apparently silently failing (I suspect stderr wasn't being written to
the log but I'm not certain).
As I say, I know that I'm referring to Windows here however this thread was
the only one we found describing "WAL archives shipped but not restored by
slave" so I've posted here in case permissions was also your issue David.
--
View this message in context: http://postgresql.1045698.n5.nabble.com/Trouble-with-replication-tp5758062p5770583.html
Sent from the PostgreSQL - general mailing list archive at Nabble.com.
