How FlexCache volumes share space with other volumes

You can have multiple FlexCache volumes in the same aggregate; you can also have regular FlexVol volumes in the same aggregate as your FlexCache volumes. To set up your system most efficiently, you should understand the way these volumes share space.

When you put multiple FlexCache volumes in the same aggregate, each FlexCache volume reserves only a small amount of space (as specified by the flexcache_min_reserved volume option—by default, 100 MB). The rest of the space is allocated as needed. This means that a “hot” FlexCache volume (one that is being accessed heavily) is permitted to take up more space, while a FlexCache volume that is not being accessed as often will gradually be reduced in size.

If you have regular FlexVol volumes in the same aggregate as your FlexCache volumes, and you start to fill up the aggregate, the FlexCache volumes can lose some of their unreserved space (if they are not currently using it). In this case, when the FlexCache volume needs to fetch a new data block and it does not have enough free space to accommodate it, a data block is ejected from one of the FlexCache volumes to make room for the new data block.

If ejected data is causing too many cache misses (as shown by the flexcache stats command), you can add more space to your aggregate or move some of your data to another aggregate.