Even more interesting details; This seems to only happen when the packets are fragmented. On 11/15/05, Ryan Castellucci <ryan.castellucci@xxxxxxxxx> wrote: > A bit more detail. > > I have the following htb classes set up... > > class htb 1:356 parent 1:4 leaf 356: prio 4 quantum 1600 rate 12800bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:357 parent 1:4 leaf 357: prio 4 quantum 1600 rate 12800bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:2 root rate 51200bit ceil 54400bit burst 64Kb/8 mpu 0b > overhead 0b cburst 1667b/8 mpu 0b overhead 0b level 7 > class htb 1:3 parent 1:2 leaf 3: prio 2 quantum 10400 rate 83200bit > ceil 88640bit burst 64Kb/8 mpu 0b overhead 0b cburst 1709b/8 mpu 0b > overhead 0b level 0 > class htb 1:614 parent 1:5 leaf 614: prio 6 quantum 1500 rate 10240bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:4 parent 1:2 rate 25600bit ceil 51680bit burst 64Kb/8 mpu > 0b overhead 0b cburst 1663b/8 mpu 0b overhead 0b level 6 > class htb 1:613 parent 1:5 leaf 612: prio 6 quantum 1500 rate 10240bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:5 parent 1:2 rate 20480bit ceil 51680bit burst 64Kb/8 mpu > 0b overhead 0b cburst 1663b/8 mpu 0b overhead 0b level 6 > class htb 1:612 parent 1:5 leaf 612: prio 6 quantum 1500 rate 10240bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:6 parent 1:2 rate 5120bit ceil 43520bit burst 64Kb/8 mpu > 0b overhead 0b cburst 1653b/8 mpu 0b overhead 0b level 6 > class htb 1:868 parent 1:6 leaf 868: prio 7 quantum 1500 rate 2560bit > ceil 43520bit burst 15Kb/8 mpu 0b overhead 0b cburst 1653b/8 mpu 0b > overhead 0b level 0 > class htb 1:869 parent 1:6 leaf 869: prio 7 quantum 1500 rate 2560bit > ceil 43520bit burst 15Kb/8 mpu 0b overhead 0b cburst 1653b/8 mpu 0b > overhead 0b level 0 > class htb 1:358 parent 1:4 leaf 358: prio 4 quantum 1600 rate 12800bit > ceil 51680bit burst 15Kb/8 mpu 0b overhead 0b cburst 1663b/8 mpu 0b > overhead 0b level 0 > class htb 1:870 parent 1:6 leaf 870: prio 7 quantum 1500 rate 2560bit > ceil 43520bit burst 15Kb/8 mpu 0b overhead 0b cburst 1653b/8 mpu 0b > overhead 0b level 0 > > and my traffic is being classified properly, however, UDP traffic is > able to exceed the ceiling rate on 1:613 and it's parents, and tops > out at about 4x whatever I set that ceil to (I tried several values, > and the UDP transfer rate always settled at about 4x the ceil). > > It manages to throttle TCP traffic just fine. ICMP exceeds the ceil > slighlty, but not enough to really worry me (about 3%) > > I've read through the documentation several times, but I don't see a > whole lot. I also tried messing with the quantum, burst, and cburst, > but that didn't really help. > > If anyone has any other ideas of what to try, i'd really appreciate > it, as I'm kinda stuck here :( > > On 11/14/05, Ryan Castellucci <ryan.castellucci@xxxxxxxxx> wrote: > > What's going on here? I'm spewing UDP traffic at this thing, and it is > > exceeding the ceil. Anyone know how to fix this? > > > > class htb 1:613 parent 1:5 leaf 613: prio 6 quantum 2560 rate 20480bit > > ceil 103360bit burst 15Kb/8 mpu 0b overhead 0b cburst 1728b/8 mpu 0b > > overhead 0b level 0 > > Sent 16591370 bytes 4159 pkt (dropped 39449, overlimits 0 requeues 0) > > rate 412384bit 6pps backlog 0b 126p requeues 0 > > lended: 887 borrowed: 3146 giants: 1748 > > tokens: -1605047 ctokens: -32828 > > > > -- > > Ryan Castellucci http://ryanc.org/ > > > > > -- > Ryan Castellucci http://ryanc.org/ > -- Ryan Castellucci http://ryanc.org/ _______________________________________________ LARTC mailing list LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
