I'm currently trying to improve the backup speed performance. I was looking at the backup job's logs, and the bptm process is is waiting thousands of times more than the bpbkar process. Does this mean that the bottleneck is on the client side, and does anyone have any ideas as to how to improve performance if this is the case?
It actual looks like the client is sending data fast enough
waiting for empty buffer = data arriving faster than the storage device can write.
waiting for full buffer = data arriving to slow. The storage device is waiting for the next block of data to write
If not done already set :
SIZE_DATA_BUFFERS = 262144 (default value for some time time in NBU).
NUMBER_DATA_BUFFERS = 256 or 512
Tuning buffers is the number one tweak when dealing with tape.
The setting must be set on all media servers.
The buffer setting is written in the activity job log for each backup job.
There may be restrictions on the number of buffers if using FTL option i Netbackup.