Hi, I've been asked to increase the queue_depth parameter for some virtualized hdisks in a HACMP cluster.
My question is about queue_depth value on the clients and on the VIOs.
If queue_depth is 8 on the VIOs can I set queue_depth to 8 also in the LPARs, or should I set a lower value?
The cluster is formed by 2 LPARs (POWER 6 servers, not sure right now if 550 or 520), double VIO for both LPARs, and the luns are from a HP EVA8400 system.
Rigth now queue depth is 3 on the client LPARs, and 8 on the 4 VIO Servers (double VIO on 2 servers).
Downtime is not a problem, as I also need to change other parameters (aio maxservers) so I will reboot.
Do you have some best practices on the queue_depth values, in VIO environment?
Thanks
My question is about queue_depth value on the clients and on the VIOs.
If queue_depth is 8 on the VIOs can I set queue_depth to 8 also in the LPARs, or should I set a lower value?
The cluster is formed by 2 LPARs (POWER 6 servers, not sure right now if 550 or 520), double VIO for both LPARs, and the luns are from a HP EVA8400 system.
Rigth now queue depth is 3 on the client LPARs, and 8 on the 4 VIO Servers (double VIO on 2 servers).
Downtime is not a problem, as I also need to change other parameters (aio maxservers) so I will reboot.
Do you have some best practices on the queue_depth values, in VIO environment?
Thanks