You may notice that your Internal database backups are increasing in the time they take to run or find that a particular operation such as a backup or restore is taking longer than it used to by a significant margin.
The issue may be tied to an issue within the Internal Database that is not being seen through the session reports or debug.log but may be reported in the PG_logs
check the pg_log directory
Win \ProgramData\OmniBack\server\db80\pg\pg_log\
Unix: /var/opt/omni/server/db80/pg/pg_log
(not to be confused with the PG_xlog!)
Here you’ll see logs generated to common IO operations. The level to which these logs are generated are located in the PostGresql.conf file but we discourage changing this unless directed by support due to the
Potential log growth with detrimental results that may occur.
In a healthy database the PG_logs (one of which is created at the start/stop of PostGres database) will contain very little information and at a glance would be only a few KB’s large.
Ex:
2015-10-06 08:42:24 EDT LOG: database system was shut down at 2015-10-06 08:32:47 EDT
2015-10-06 08:42:24 EDT LOG: database system is ready to accept connections
2015-10-06 08:42:24 EDT LOG: autovacuum launcher started
2015-10-06 08:42:24 EDT FATAL: the database system is starting up
HOWEVER, under conditions where there may be some database inconsistencies or other irregularity the PG_log may get very large and this would be indicative of a problem in the database that would need support intervention.
In general these types of inconsistencies are known to support and a script can be delivered to fix the issue. Some other issues that cause excessive log growth have been addressed in latest patches.
So, to recap. Small PG_logs are ok but large PG_logs may be indicative of an issue needing investigation. Open a case and support will provide further assistance.