[Asterisk-Users] Asterisk Crashing (high load issues)

[Kyle Hagan]

The problem seems to be related to the number of calls being dialed,
and not the number of agents logged in.
Is there any configuration for the Digium cards that would make it work
better? The older server is not crashing with
50+ agents logged in, the load is just high.

We were planning on getting the DS3 card digium is supposed to release
soon. But it seems the problem is dialing 100+ lines
at a time. Will the DS3 card help asterisk handle making the calls? If
not what use is the DS3 card, right?

Today we are going to try Fedora Core 1 x86_64 and Mandriva 2005 x86_64
to see if it works better... Any ideas?

Here is the actual messages file:

  9 15:18:12 xeonAsterisk kernel: zaptel Disabled echo canceller because 
of tone (rx) on channel 12
Nov  9 15:18:20 xeonAsterisk kernel: zaptel Disabled echo canceller 
because of tone (rx) on channel 39
Nov  9 15:19:12 xeonAsterisk kernel: BUG: soft lockup detected on CPU#3!
Nov  9 15:19:12 xeonAsterisk kernel:
Nov  9 15:19:12 xeonAsterisk kernel: Modules linked in: md5 ipv6 
parport_pc lp parport autofs4 i2c_dev i2c_core rfc
omm l2cap bluetooth sunrpc pcmcia yenta_socket rsrc_nonstatic 
pcmcia_core dm_mod video button battery ac uhci_hcd e
hci_hcd wct4xxp(U) zaptel(U) crc_ccitt shpchp tg3 floppy ext3 jbd 
megaraid_mbox megaraid_mm sd_mod scsi_mod
Nov  9 15:19:12 xeonAsterisk kernel: Pid: 4136, comm: asterisk Not 
tainted 2.6.11-1.1369_FC4smp
Nov  9 15:19:12 xeonAsterisk kernel: RIP: 0010:[<ffffffff8013f198>] 
<ffffffff8013f198>{__do_softirq+104}
Nov  9 15:19:12 xeonAsterisk kernel: RSP: 0018:ffff8100033b3f68  EFLAGS: 
00000202
Nov  9 15:19:12 xeonAsterisk kernel: RAX: ffffffff80510800 RBX: 
0000000000000002 RCX: 00000001003ac615
Nov  9 15:19:12 xeonAsterisk kernel: RDX: 0000000000000000 RSI: 
ffff810002c28680 RDI: 0000000000000002
Nov  9 15:19:12 xeonAsterisk kernel: RBP: ffff81007597fe18 R08: 
0000000000000000 R09: 0000000000000000
Nov  9 15:19:12 xeonAsterisk kernel: R10: 0000000000000000 R11: 
000000000000008c R12: ffffffff80513980
Nov  9 15:19:12 xeonAsterisk kernel: R13: 0000000000000180 R14: 
ffffffff80557a80 R15: 000000000000000a
Nov  9 15:19:12 xeonAsterisk kernel: FS:  0000000040aea960(0063) 
GS:ffffffff80510800(0000) knlGS:0000000000000000
Nov  9 15:19:12 xeonAsterisk kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 
000000008005003b
Nov  9 15:19:13 xeonAsterisk kernel: CR2: 0000000000ab4a7c CR3: 
0000000077d12000 CR4: 00000000000006e0
Nov  9 15:19:13 xeonAsterisk kernel:
Nov  9 15:19:13 xeonAsterisk kernel: Call Trace: <IRQ> 
<ffffffff8013f260>{do_softirq+48} <ffffffff8010f419>{apic_ti
mer_interrupt+133}
Nov  9 15:19:13 xeonAsterisk kernel:         <EOI> 
<ffffffff80189843>{fget+83} <ffffffff80189829>{fget+57}
Nov  9 15:19:13 xeonAsterisk kernel: 
<ffffffff8019dd14>{sys_poll+452} <ffffffff8019d590>{__pollwait+0}
Nov  9 15:19:13 xeonAsterisk kernel: 
<ffffffff80111c47>{syscall_trace_leave+55} <ffffffff8010ebf6>{tracesys+
209}
Nov  9 15:19:13 xeonAsterisk kernel:
Nov  9 15:19:13 xeonAsterisk kernel:
Nov  9 15:19:13 xeonAsterisk kernel: Call Trace: <IRQ> 
<ffffffff80162344>{softlockup_tick+132} <ffffffff80113481>{t
imer_interrupt+481}
Nov  9 15:19:13 xeonAsterisk kernel: 
<ffffffff801625ec>{handle_IRQ_event+44} <ffffffff8016271d>{__do_IRQ+253
}
:



Kyle

Matt Florell wrote:

>I would say your best bet is to change your system into a distributed
>dialing system. We did this with Vicidial and have installations on
>multiple servers with over 100 agents all working off of the same
>lists and campaigns. A distributed system will also allow for more
>redundancy and less total downtime if one server goes down.
>
>We noticed the same kind of limitations you are and now do a max of 40
>agents per server, and when we need more capacity we just add another
>server.
>
>MATT---
>
>On 11/10/05, Kyle Hagan <info at quadrasoftware.com> wrote:
>  
>
>>Kyle Hagan wrote:
>>
>>    
>>
>>>We purchased a new Dual Xeon 3ghz, 2gb ram to upgrade our 3ghz Pentium
>>>1gb ram, that has been having load issues due to our growing company.
>>>
>>>We are having problems... We use a predictive dialer that we custom
>>>programmed in perl. It basically drops, moves, files into the callout
>>>directory and uses queues to transfer to agents when someone picks up.
>>>      
>>>
>>Oh, we are running HEAD version.
>>
>>
>>Kyle
>>_______________________________________________
>>--Bandwidth and Colocation sponsored by Easynews.com --
>>
>>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
>>
>>    
>>
>_______________________________________________
>--Bandwidth and Colocation sponsored by Easynews.com --
>
>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
>
>  
>