[Asterisk-Users] Bad Bad Performance; Max 20 Calls on Quad Proc?

[Asterisk]

We're running a 6400 (cat /proc/cpuinfo only returns a single 733 Mhz 
CPU but I'm sure it's a dual cpu machine. Perhaps it's not running SMP) 
with 83 registered SIP phones. We are running ulaw/alaw, with a TE405P 
linked to a E1 EuroISDN. Oh, and we've only got 512MB memory. We have 4 
call queues, 45 agents and all agent calls are recorded (but not mixed).

The OS is fedora core 1.

We do not have any performance problems (although we are using swap too 
much - more memory would be good). We have a 3ghz 4gb memory machine 
waiting to be brought on line.

So - we are simply not having the same performance problems that you 
seem top be having.

I've included a shot of top, but this is with no-one online (everyone's 
gone home!)

  22:14:47  up 74 days, 10:17,  2 users,  load average: 0.00, 0.00, 0.00
56 processes: 55 sleeping, 1 running, 0 zombie, 0 stopped
CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
            total    0.1%    0.0%    0.3%   0.0%     0.0%    0.0%   99.4%
Mem:   514660k av,  499512k used,   15148k free,       0k shrd,   44912k 
buff
         74636k active,              92196k inactive
Swap: 1044216k av,   10272k used, 1033944k free                   90904k 
cached


Julian.


Matthew Boehm wrote:
> Here is the setup: Dell 6250, Quad Proc P3 500Mhz. Digium Single Span T1 
> card. System has 71 sip peers/users. All calls are G729; we have 10 
> licenses. All calls follow this path: UA -> Asterisk -> Digium PRI -> 
> Class4/5 switch. The switch dictates if it should go out local PRI for 
> local termination or out PRI to our Cisco AS5300 for LD termination.
> 
> Our biggest problem is echo. Sometimes there is none, sometimes there is 
> A LOT, sometimes there is a little. Sometimes it happens on local 
> PRI->PRI calls, sometimes it happens on Voip->voip calls.
> 
> I've been watching 'top' again.  I witnessed the CPU usage on Asterisk 
> jump to and stay at 47% with 4 calls up. Then it fell back to 1% with 
> those same 4 calls about 10 seconds later.
> 
> I'm very dissapointed in this. Is this what I can expect? I still can't 
> grasp why asterisk is not using the other 3 processors to better its 
> performance.
> 
> But then again, reading Digium's new benchmarks for their new card this 
> sorta fits. They benched a dual 3Ghz Xeon and got 120 G729->PRI calls.
> 
> Another thing is "QUALIFY" problems.  All our SIP users are on T1's that 
> are directly connected to our network. All our customers' T1 come 
> straight to our Cisco 7206. Our asterisk box is on a Catalyst switch 
> also connected to the router. All but 15 of the 71 peers we have, have 
> qualify turned on; Mostly because I need the NAT keep-alive. All day 
> long I will get "Too Lagged" or "Unreachable" on UAs then several 
> seconds later they come back. Could this be overloading cpu? Seems this 
> could be put onto another proc.
> 
> Some of that is rant; some is pure concern.  Is anyone else experiencing 
> similiar stuff?
> 
> -Matthew
> 
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users
> To UNSUBSCRIBE or update options visit:
>   http://lists.digium.com/mailman/listinfo/asterisk-users
> 
>