dont-use-fsync real world impact

[Jure Sah]

Hi,

The documentation clearly states that using the dont-use-fsync option is
not recommended.

I am wondering what is the realistic impact of this? As I understand the
kernel commits dirty pages to disk every 30 seconds by default, and this
is configurable. Wouldn't this mean that at worst 30 seconds worth of
leases are lost?

The leases file is in most cases relatively tiny (under 1 MB), and could
easily fit in system storage cache. However the fact that it gets
fsynced on every commit means that the performance of the DHCP server is
capped at whatever the performance of the physical storage is. While
fast storage options exist, this doesn't make the DHCP server
future-proof as a solution, which is a problem.

>From the past correspondence from the mailing list archive I surmise
that people usually work around this by using hardware cache that does
not obey fsync, which simply offloads the problem from the kernel to the
cache controller and only superficially solves the problem. It however
hints towards the view that perhaps not using fsync is not all that bad,
if we are talking about a typical professionally hosted server.

LP,
Jure