DHCP Queue

[Brad Dameron]

Also your logging could be a issue. I posted to the list how to set logging to cache more which can increase performance two-fold as well. This is for syslog and syslog-ng.
 
Brad Dameron
Clearw're
Engineering

________________________________

From: dhcp-users-bounces at lists.isc.org on behalf of Blake Hudson
Sent: Thu 2/12/2009 11:39 AM
To: Users of ISC DHCP
Subject: Re: DHCP Queue


And if you can't run the newest version of DHCP... If your SAS controller has a write cache, enable it. The bottleneck in your DHCP server is likely disk. If you can remove that, you could probably serve a few thousand lease per second with your CPU.

-Blake

-------- Original Message  --------
Subject: Re: DHCP Queue
From: Brad Dameron <Brad.Dameron at clearwire.com> <mailto:Brad.Dameron at clearwire.com> 
To: Users of ISC DHCP <dhcp-users at lists.isc.org> <mailto:dhcp-users at lists.isc.org> 
Date: Thursday, February 12, 2009 11:26:07 AM


	You can increase the lease per second substantually by going with 3.1.x and using the hash adjustment I have posted on this list before. I have gotten it into the 400 lps range with this. And I also tested with DHCPperf tool as well as some other in house generated perl scripts. Search the archive for my name.
	 
	Brad Dameron
	Clearw're
	Engineering
	 

________________________________

	From: dhcp-users-bounces at lists.isc.org on behalf of T MANIKANDAN-PKXR74
	Sent: Wed 2/4/2009 9:01 AM
	To: Users of ISC DHCP
	Subject: RE: DHCP Queue
	
	
	Hello can any one help to my query mentioned below.
	 
	Regards
	Mani

________________________________

	From: dhcp-users-bounces at lists.isc.org [mailto:dhcp-users-bounces at lists.isc.org] On Behalf Of T MANIKANDAN-PKXR74
	Sent: Tuesday, February 03, 2009 7:30 PM
	To: dhcp-users at lists.isc.org
	Subject: DHCP Queue
	
	
	Hi,
	 
	   We have a Network element AAA where it sends 250 clients per second to aquire a Lease from ISC DHCP server. and I have seen that a 4 way quad core AMD opteron processor/ SAS I/O drive with RHEL 5 and ISC DHCP 3.0.5 can handle upto 125 lease per second ( Result achieved from Nominum DHCPperf tool).
	 
	   With total 135000 clients in my network it takes around 18 min to complete the lease ( Time taken= (135000/125)/60 = 18 Minutes), When my AAA network element which is used for authentication and other purpose finishes the job in approx 8 Minutes, while my DHCP server take 18 Minutes to complete the Job.
	 
	   Say at the end of 3 seconds, DHCP server gets a hit of 750 (250+250+250) Discover request while only 375 (125+125+125) is being allocated an ip address and balance 375 will be in a queue similarly..... 
	  at the end of 500 seconds, my DHCP server would have got aprox 135000 discover request and it has processed only approx 65000 request and balance 75000 will be pending in the queue.
	 
	My queries
	 
	How long the client request will remain in queue 
	what will be the maximum attempt the client will make to obtain the lease.
	What will be the state of clients which are there in queue.
	Is there any other Best solution in which my DHCP server can handle all 250 clients at a second 
	Can i have 2 DHCP server in active active state handling different scopes so that all 250 request are handled and there is no queue if so how the lease file will be maintained.
	 
	 
	  Please help me understanding how the DHCP server handles this situation
	 
	Regards
	Mani
	
________________________________


	_______________________________________________
	dhcp-users mailing list
	dhcp-users at lists.isc.org
	https://lists.isc.org/mailman/listinfo/dhcp-users


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20090212/48a136a5/attachment.html>