Problem
After SCO ARCserve has completed a backup the SCO UNIX system becomes very slow while ARCserve re-indexes its database.
Cause
Normally any backup job is scheduled during off-peak times. SCO ARCserve will re-index its database after completion of a backup. If an SCO ARCserve backup has failed and is automatically rescheduled to a "peak" time, after the rescheduled backup has completed ARCserve will re-index its database in a low-priority, low system resource mode. On some systems this re-indexing will cause system resources to be over taxed despite this task being assigned low-priority.
Solution
Follow any or all of the suggestions below to allow ARCserve to function in a non-intrusive way.
1. Make sure the SCO ARCserve database is purged and pruned to
a manageable size. ARCserve will complete indexing in a shorter
amount of time if its database is smaller, however pruning and
purging will not reduce the amount of resources used during
indexing.2. Schedule "autopilot" jobs so that "reschedule job" is
disabled if a backup job fails (tape labeled incorrectly,
network down or client down, and so on). If an "autopilot"
job fails and is rescheduled, typically the rescheduled time
is 12 hours later. As an example, if the original backup time
was 1:00 A.M. and that backup failed ARCserve to attempt to
complete that failed backup at 1:00 P.M., typically a peak time.3. Schedule incremental backup jobs during weekdays and full backups on typical non-peak times, such as a weekend day. Typically the size of incremental backup jobs are about 10% of full backups.
See also
man as_backup (ARC), as_dbmgr (ARC); TA100115 "What is the ARCserve database and how do I maintain it?"