Troubleshooting poor backup performance

When a node is performing very poor (for example, the backup is running for 8,5 hours and the TSM server only received 150 Mb of backup data)
That time we want to check for the following:
  • The NIC settings; half/full duplex, auto sense, 10/100, etc.
  • Check/compare the network speed against the aggregate network speed
  • Try to turn compression on or off on the client (compression on the client requires more processor cycles locally)
  • Try a tuned option set (or set this in dsm.opt) with different resource utilization values
  • Try a tuned management class (or set this in dsm.opt) with different serialization ((shr)static vs. (shr)dynamic)
Previous
Next Post »