How the default queue works

Understanding how the default priority is used helps you create the optimal priority policy for your storage system.

Any volume that does not have a priority assigned is in the default queue. If you have not assigned a priority to any volume on your system, then all of your volumes are in the default queue, and requests to all volumes are given equal priority.

When you assign a priority to any volume, it is removed from the default queue. Now, requests to that volume are assigned priorities relative to requests for the default queue. But all of the volumes in the default queue share the resources allocated to the default queue. So if you assign priorities to a few volumes and leave the rest in the default queue, the results may not be as you expect.

For this reason, once you assign a priority to any volume, you should assign a priority to all volumes whose relative performance you want to control.

For example, you have 30 volumes on your system. You have one volume, highvol, that you would like to have faster access to, and one volume, lowvol, for which fast access time is not important. You assign a priority of VeryHigh to highvol and VeryLow to lowvol. The result of these changes for the highvol volume is as expected: when the system is under load, accesses to the highvol volume are given a higher priority than for any other volume. However, accesses to the lowvol volume may still get a higher priority than accesses to the volumes that remain in the default queue (which has a Medium priority). This is because all of the 28 volumes remaining in the default queue are sharing the resources allocated to the default queue.