FAST VP Pool full condition-need to move back data

you can configure Relocation Rate to spped up data movement.

The Relocation Rate is a quality of service (QoS) setting for FAST VP and affects the ―aggressiveness‖ of data movement requests generated by FAST VP. This aggressiveness is measured as the amount of data that will be requested to be moved at any given time, and the priority given to moving the data between pools. The relocation rate can be configured to be between 1 and 10, with 1 being the most aggressive. The default is 5.

see your current Relocation rate using command:

symfast -sid xxx list -control_parms

Set relocaton rate using command:

symfast -sid 207 set -control_parms --vp_reloc_rate xx

This is the relocation rate, but what I wanted to know was, if setting the PRC from 1% to 20% will move data from the sata pool to the other pools. Looks like it does move the data when PRC is increased, the pool allocation is now down to 96% from 99%.

If the PRC is increased, causing a thin pool to be within the PRC limit, FAST VP will not automatically start moving data out of the pool.

The PRC affects only the ability of FAST VP to move data into a pool.

deykau,

I believe debmig is correct. (with a small correction, causing the thin pool not to be within PRC limit.)

The change in pool allocation from 99% to 96% that you noticed may be due to diferent factors such as performance movements of FAST VP.

Santhosh, so you mean that to comply to the PRC value on the pool, data movement will not happen out of the pool? Data movement due to performance does not seem very likely here, because it would not have moved to the SATA pool at the first place if it had been for performance. The pool became 99% and stayed there for quiet sometime before I changed the PRC, and as soon as I changed the PRC I could see some of the data moving out of the pool. I dont know, but it still seems very likely that data moved out of the pool to make it compliant to PRC.

deykau,

PRC determines how much tdev extents can be moved into the pool by FAST VP.

can you paste the output of the below command from the box where you have encountered the issue:

symfast -sid xxx list -control_parms

Alas there is no way to check the FAST VP extent movement like in FAST, but we can still determine which extent of the devices are currently residing at which pool.

Deykau,

Yes, I meant the same, the PRC value only affects the ability of FAST VP to move data into a pool.

If PRC is increased for a pool causing the pool to be out of compliance with the PRC limit, FAST VP will not automatically move data out of the pool.

That could be a coincidence that as soon as you changed the PRC limit, FAST VP might have started the performance movements (movements determined by the performance requirements of the extents) and it could have appeared to you as if they were caused by changing the PRC to comply with it.

deykau, are you still having any doubts pertaining this. If not it will be nice if you mark this discussion as answered !