Avoid filling zpools beyond 80% of their capacity

Keep pool space under 80% utilization to maintain pool performance. Currently, pool performance can degrade when a pool is very full and file systems are updated frequently, such as on a busy mail server. Full pools might cause a performance penalty, but no other issues. If the primary workload is immutable files (write once, never remove), then you can keep a pool in the 95-96% utilization range. Keep in mind that even with mostly static content in the 95-96% range, write, read, and resilvering performance might suffer.

•    Issues specific to 80% full zpool are:
o    If the zpool is fragmented and has less free space available, then it will take longer and require more CPU cycles in the kernel to find a suitable block of free space for each write.  This results in lower write performance if the zpool has less than 20% free space. This issue is addressed by the following document:

SunAlert: ZFS(zfs(1M)) filesystem(5) Performance May Drop Significantly if the ZFS Pool Becomes Full (Doc ID 1019947.1)

Bug 15418573 / 6596237 - Stop looking and start ganging

Even with the fix, write performance still degrades as the pool becomes more fragmented and running with less than 20% free. There is another bug open to address the short coming.

Bug 15702274 / 7026795: Stop looking and start ganging, really
The fix is still being worked by ZFS sustaining.

Comments

Popular posts from this blog

BMCs and the IPMI Protocol

Logical Domains Reference Manual

Understanding How ZFS Calculates Used Space