Watching performance on a DHCP Server

[Simon Hobson]

Blake Hudson wrote:

>The pros of fsync are that I can say I will never lose a lease 
>(assuming my disks don't die, RAID controller doesn't die, file 
>system doesn't corrupt, RAM doesn't go bad, server doesn't hit the 
>bricks, etc).


Actually, for most of those, you still won't lose  a lease ! The 
point is, that if the write doesn't happen, then the client isn't 
offered the lease.

>From a practical perspective, the thought of losing 5 seconds of 
>leases in server lockup is a non-issue to me (and probably others).

Actually, I would suggest that losing leases IS a big deal and can 
cause some ongoing problems (mostly with bad clients).

>If I had the choice of penalizing every request, versus penalizing a 
>few requests in several years time I would choose the later.

That's the tradeoff.

I can't see any problem including a config option, provided it 
doesn't appear in the config file by default, AND there's suitable 
warnings in the docs (like anyone reads those !)

IIRC someone on this list did a test a while back - of the form 
"throw loads of requests at the server and yank the plug, see what 
happens when it comes back up". The ISC server had a lower throughput 
but correctly recovered with a knowledge of what leases it had given. 
A certain popular product from Redmond got through loads more client 
requests (a couple of orders of magnitude IIRC) but lost loads of 
them after the reboot.