Watching performance on a DHCP Server

Frank Bulk - iNAME frnkblk at iname.com
Sun Feb 17 05:29:53 UTC 2008


Thanks for your response.  We have separate DHCP appliances, which syslog
locally and across the network.  But our volumes are so low, it's only hard
on the external syslog receiver which is busy enough doing many other
things.

And we do have fsync off for most of our syslogged info.

Frank

-----Original Message-----
From: dhcp-users-bounce at isc.org [mailto:dhcp-users-bounce at isc.org] On Behalf
Of John Hascall
Sent: Saturday, February 16, 2008 4:46 PM
To: dhcp-users at isc.org
Subject: Re: Watching performance on a DHCP Server 


> At these volumes/rates, are people syslogging DHCP traffic to disk?

We are.  However, our platform (NetBSD) does not fsync after
non-kernel calls to syslog.  If you are using a platform where
syslog (e.g., [at least some variants of] Linux) fsync's by
default you might wish to disable that as it can become a
bottleneck (look for "-file" in "man syslog.conf").

Also, our syslog is on a different disk (a RAID 0)
than our leases (a RAID 1).

John


> Frank
>
> -----Original Message-----
> From: dhcp-users-bounce at isc.org [mailto:dhcp-users-bounce at isc.org] On
Behalf
> Of sthaug at nethelp.no
> Sent: Tuesday, February 12, 2008 3:49 AM
> To: dhcp-users at isc.org
> Subject: Re: Watching performance on a DHCP Server
>
> > You may want to review the thread from the beginning. My network
> > currently has 10,000+ DHCP clients (and I plan on accommodating double
> > that within the lifetime of this server). I have a beefy server (4x
> > 3.0GHz Xeon, 2x 15k RAID1) and it was only able to reliably handle 10 to
> > 20 4-way discover handshakes a second, 2-way handshakes were maybe
> > double or triple those numbers. When pounded by DHCP requests, it's
> > possible that even less are processed in a timely manner due to
> > collisions, timeouts, etc.
>
> I really don't understand the numbers people are quoting here. We use
> a Dell 1850 with one 3.2 GHz CPU, one Gig of memory and two SCSI disks
> in battery backed hardware RAID 1. We use this to serve 100K customers
> with 24 hour leases. No problems whatsoever, and we feel we have plenty
> of room to grow.
>
> (Some calculations: Clients are expected to renew after half the lease
> time has passed. So we can expect 100.000 renewals in 43200 seconds, or
> a little over 2 renewals per second. Monitoring the actual DHCP traffic
> on the server shows numbers which are consistent with this. We have run
> with considerably shorter lease times - down to one hour in connection
> with planned network reconfigurations - and again the increased DHCP
> traffic has been no problem whatsoever.)
>
> Steinar Haug, Nethelp consulting, sthaug at nethelp.no
>
>
>





More information about the dhcp-users mailing list