[rabbitmq-discuss] Huge latency in Linux, compared with Leopard

Carl Trieloff cctrieloff at redhat.com
Sun Sep 28 02:19:59 BST 2008


> - kernel settings matter, but not as much as one would think. My server
> runs at 250 HZ & voluntary preemption, whereas the laptop runs with full
> preemption at 300 HZ - however both exhibit very similar symptoms, and I
> strongly suspect you'd see the same at 1000 HZ or with the RT kernel
> (still need to try that one). Keep in mind that different distributions
> have patched kernels to varying degrees (especially RedHat) and that the
> relatively new CFQ CPU scheduler (new in 2.6.23 IIRC) had a lot of
> performance oddities since its introduction. My understanding from
> following the kernel list most of these should be fixed in the current
> 2.6.26 kernel, however the variance between the average (~1ms) and max.
> latency (40ms) is IMHO just way too big for an occasional mis-schedule so
> something else must be wrong. Besides we all get the same 40ms penalty so
> that is a good sign that it's not the kernel scheduler per se.

I'll add one note, the Red Hat MRG RT kernel gives you a max scheduler 
latency of ~20us. The
Linux TCP stack can run with max latency sub 1ms with both 2.6.18 RHEL5 
and MRG RT
2.6.24. most of your issues are most likely nagle.

Carl.




More information about the rabbitmq-discuss mailing list