Best way to benchmark SLOG

We purchased a couple of Optane’s and Radians so that we can install them into a Truenas. We use truenas for NFS VHD sharing, so want to run a benchmark with sync writes (as sync is enabled).

The idea is to test:
a) Without a dedicated SLOG
b) Test with a Radian
c) Test the Optane

Can anyone suggest the best way to run these tests? The initial idea was the run a dd command from a running VM which is hosted on the shared pool. That way we can determine the actual write speeds with or without a dedicated SLOG. I have seen some other threads where people seem to use diskinfo which also appears to run on the Truenas appliance directly. This seems counter productive as it would not be a true reflection of the speed impact over NFS sync writes.

Any tips? We’re new to benchmarking disks so any help will do

Ideas and data points in the old forum:

1 Like