Hey It's a bug in TC, upgrade to a newer version and it should be fixed. I had the same problem on a dual xeon 2.4 ghz with crashed with 100 classes loaded. Did'nt understand why... Regards Matias Bjørling ----- Original Message ----- From: "Harkisoon Sharan (CCI-Gainesville)" <Sharan.Harkisoon@xxxxxxx> To: <lartc@xxxxxxxxxxxxxxx> Sent: Monday, May 19, 2003 4:21 PM Subject: [LARTC] Problem creating a lot of classes Hello, I'm looking for a little advice. The following example works great for me on a small scale test, but I'm trying to implement this on a larger scale for a /22 address space, where every user is commited 1mbit of bandwidth. Unfortunately, when I try to create over 600 classes, I get a kernel panic (the threshold is somewhere between 600-750 for me). I am using a Dell Poweredge 2450 with 1gig of ram, dual 733mhz, with a standard Redhat9 install/kernel. Any suggestions, or alternative configurations would be appreciated. Thanks, Sharan Harkisoon Here's the script: #!/bin/sh ingress=eth0 egress=eth1 downstream=1mbit upstream=1mbit #clear shaping on interfaces tc qdisc del dev $egress root tc qdisc add dev $egress root handle 1: htb default 1 tc class add dev $egress classid 1:1 htb rate $downstream tc class add dev $egress classid 1:2 htb rate $downstream tc class add dev $egress classid 1:3 htb rate $downstream .. #ommitted .. tc class add dev $egress classid 1:1021htb rate $downstream tc filter add dev $egress protocol ip prio 1 u32 match ip dst 24.136.40.2 flowid 1:1 tc filter add dev $egress protocol ip prio 1 u32 match ip dst 24.136.40.3 flowid 1:2 tc filter add dev $egress protocol ip prio 1 u32 match ip dst 24.136.40.4 flowid 1:3 .. #ommitted .. tc filter add dev $egress protocol ip prio 1 u32 match ip dst 24.136.43.254 flowid 1:1021 _______________________________________________ LARTC mailing list / LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
