Re: [LARTC] Swap size

2006-03-22 Thread Anton Glinkov
If the machine will be used only for squid, I suggest you create no more than 1GB of swap, just in case the physical memory runs out. You can play with the cache_mem setting in squid.conf and see what is the maximum value with which it doesn't use swap-space (after all what you seek is cache

Re: [LARTC] Swap size

2006-03-22 Thread Tomas Simonaitis
I agree, 1-2GB of swap is good choice from my experience. I don't think you should run without swap at all - there usually is something OS can push to swap and free some ram. On Wednesday 22 March 2006 17:56, Anton Glinkov wrote: If the machine will be used only for squid, I suggest you create

Re: [LARTC] Swap size

2006-03-22 Thread William L. Thomson Jr.
On Wed, 2006-03-22 at 19:19 +0200, Tomas Simonaitis wrote: I agree, 1-2GB of swap is good choice from my experience. I don't think you should run without swap at all - there usually is something OS can push to swap and free some ram. Well it can't hurt in the sense that it's not like hard

[LARTC] no transmit

2006-03-22 Thread Larry Brigman
In trying to configuring classful queing using some of the examples from the lartc howto, I am getting: qdisc cbq 1: rate 10Kbit (bounded,isolated) prio no-transmit Here are my two lines to get started. tc qdisc add dev eth4 root handle 1: cbq bandwidth 100Mbit avpkt 1000 cell 8 tc class add

[LARTC] HFSC and default qdisc backlog

2006-03-22 Thread James Nelson
Thanks for all of your help Patrick! Just so I'm clear. If hfsc at the class level shows no overlimits and no packet dropps, then hfsc is not effecting my traffic any different (from a throughput perspective computational computer slowness aside) then if i had no traffic shapping in

Re: [LARTC] HFSC and default qdisc backlog

2006-03-22 Thread Jason Boxman
On Wednesday 22 March 2006 19:13, James Nelson wrote: Thanks for all of your help Patrick! Just so I'm clear. If hfsc at the class level shows no overlimits and no packet dropps, then hfsc is not effecting my traffic any different (from a throughput perspective computational computer

Re: [LARTC] HFSC and default qdisc backlog

2006-03-22 Thread Patrick McHardy
Jason Boxman wrote: On Wednesday 22 March 2006 19:13, James Nelson wrote: Thanks for all of your help Patrick! Just so I'm clear. If hfsc at the class level shows no overlimits and no packet dropps, then hfsc is not effecting my traffic any different (from a throughput perspective