On 10/1/15 8:04 PM, Joseph Kregloh wrote:
> In either case you are still "bottlenecked" by the speed of the write
> from RAM to the zpool. Now for a small database with not many writes a
> ZIL would be awesome. But on a write heavy database you will be
> acknowledging more writes because of the ZIL that what you are
> physically able to write from RAM to zpool, thereby degrading performance.
Unless ZFS lies about fsync, you'll have to wait for writes to go
somewhere, either the main pool or the ZIL. Because the ZIL is
effectively write-only, having a ZIL could significantly reduce fsync
latency because it's a) only writing and b) writing sequentially.
Essentially, it's the same idea as having a separate pg_xlog partition
(except that depending on your setup there may well be pg_xlog read
activity as well).
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Experts in Analytics, Data Architecture and PostgreSQL
Data in Trouble? Get it in Treble! http://BlueTreble.com
