the archive command is a cp command: 'test ! -f /bigpart/local/incoming/%f && cp %p /bigpart/local/incoming/%f && chmod a+rx /bigpart/local/incoming/%f'
Your primary and secondary instances/db clusters are on the same server?
No, two servers: the temp partition (/bigpart/local/incoming) is on the master and is shared via NFS to the standby (/mnt/db1_incoming_wals). On the standby server the partition is mounted in rw mode and the postgres user can read and write (and delete) on the partition (verified now).
I modified the command in the recovery.conf as follow, but no changes and no entry in the log file: archive_cleanup_command = '/usr/pgsql-9.3/bin/pg_archivecleanup /mnt/db1_incoming_wals %r 2>>/tmp/archive_cleanup.log'