[Asterisk-Users] Dual T400P, SMP, performance issues
Alex Zarubin
ZAlex at Webley.COM
Mon Jun 9 17:08:12 MST 2003
Hi,
We are trying to validate Asterisk as a media gateway PRI <-> SIP with two
T400P (8 T1s) per box. The first
experience with BOX1 (Compaq, 2.53 GHz, 1 Gb RAM) and just one T400P was
encouraging - on the load
test with 3 T1s worth of calls we had on average 75% idle CPU.
Not so with BOX2 (Dell, single 2.6 GHz Xeon, 1 Gb RAM, 2 T400P) and BOX3
(Dell, dual 2.6 GHz Xeon,
2 Gb RAM, 2 T400P, asterisk/zaptel is built with SMP support).
On the similar load test (as with the BOX1) BOX2 was showing 0% idle CPU 70%
of the time. Just 3 T1s
out of 8.
On the load test with just 2 T1s BOX3 was very close to 0% idle on CPU0,
CPU1 was at 95% idle.
The process ksoftirqd_CPU0 was close to the top of the 'top', with
/proc/interrupts showing tor2 related
numbers growing very fast. We had 2 T1s plugged into the first T400P board,
with nothing going into the second,
but the number of interrupts for the both boards was growing at the same
pace. Here are the interrupts
(after the box reboot, so they are not that big as they were) - do they look
OK?
CPU0 CPU1 CPU2 CPU3
0: 122556 0 0 0 IO-APIC-edge timer
1: 4 0 0 0 IO-APIC-edge keyboard
2: 0 0 0 0 XT-PIC cascade
5: 0 0 0 0 IO-APIC-level usb-ohci
8: 1 0 0 0 IO-APIC-edge rtc
12: 20 0 0 0 IO-APIC-edge PS/2 Mouse
14: 23 0 2 0 IO-APIC-edge ide0
20: 516930 0 0 0 IO-APIC-level tor2
24: 516524 0 0 0 IO-APIC-level tor2
28: 10600 0 0 0 IO-APIC-level eth0
29: 4837 0 0 0 IO-APIC-level eth1
30: 24831 0 0 0 IO-APIC-level aacraid
NMI: 0 0 0 0
LOC: 122430 122429 122429 122428
ERR: 0
MIS: 0
Not sure what went wrong. Any suggestions on how to work with 2 T400P in a
box (without hurting performance)
and how to get advantage of SMP for Asterisk would be appreciated.
Any known Linux kernel related issues (2.4.20-13.7smp #1 SMP for BOX3 )?
Thank you.
Alex Zarubin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20030609/11a3bfd0/attachment.htm
More information about the asterisk-users
mailing list