Logs Produced During Synthetic Backups

When a synthetic backup is scheduled, the scheduler (bpsched) starts program bpsynth to manage the synthetic backup process. (An administrator may also run bpsynth directly from the command line.) bpsynth plans how the synthetic backup will be built from the previous backup images.

bpsynth then calls program bpcoord to coordinate reading the necessary files from each of the component images, one image at a time. (bpcoord cannot be run directly from the command line.)

bpsynth passes information to programs bptm and bpdm to cause tape and disk images to be read or written. Catalog information is managed using bpdbm. Each of these programs has a debug log file in the logs directory. If problems occur with synthetic backups, the following debug logs are required to diagnose the problem. See the NetBackup Troubleshooting Guide for UNIX and Windows for more information.

On the master server: bpsynth, bpcoord, bpdbm, bpsched, bpcd.

On the media server(s): bptm (tape images), bpdm (disk images), bpcd.

Note that several media servers may be involved if the component images are on different nodes.

Only one bpsnth/bypcoord pair runs at a time for any policy. This means to make a synthetic backup for several clients in a policy, the clients will be backed up one after another. Also, if the policy uses multiple streams, each stream will be synthesized one after the other.

Multiple streams and multiplexing provide parallelism that speed up the total traditional backup time. Multiplexed and multi-stream images are processed sequentially (one client or one stream at a time) during a synthetic backup. The sequential processing of these images during a synthetic backup may be acceptable if the synthetic backups take place during periods of relative inactivity on the media server. Better performance can be achieved if the component images are on disk (disk or disk-staging storage units).