Symptoms
Upgraded to v9.1.
nsrclone started with a workflow uses lots of memory, times out, and the job fails.
nsrclone from CLI => working OK.
Log shows:
Unable to setup direct save with server nsrserver: busy.
NSR server nsrserver: busy
waiting 30 seconds then retrying
Read devices show 'ready for reading, idle' write devices show 'writing, idle' and the nsrclone processes consume more memory than in 9.0.x.
Problem most likely to show if the clone query time is greater than one day.
Cause
Server resource issue due to new clone features in v9.1.
Resolution
Create file:/nsr/debug/nsrcloneconfig (Or on Windows: \Program Files\Legato\nsr\debug\nsrcloneconfig )Note no file extension.Add the following lines: max_total_dd_streams=256 ams_enabled=yes ams_slice_size_factor=31 ams_preferred_slice_count=0 ams_min_concurrent_slice_count=1 ams_max_concurrent_slice_count=20 ams_force_multithreaded=yes and "Disable RPS clone: No" This is set the NetWorker Management Console, in the configuration tab on the left, there will be a check box to Disable RPS cloning.Note:max_total_dd_streams=256This depends on the DDs involved in the clone read/write operations. Check the specification of the DD models and adjust this figure accordingly. 256 is too high for smaller model DDs.
