Page 1 of 1
Vicibox server performance with SIP
Posted:
Fri Apr 09, 2010 9:15 am
by dimorry
Hi Fellows,
I have a 100% dial out operation, using manual dial (for now) running on a dell poweredge Dual 3.0 GHZ processor, 4 GB RAM and 10k SCSI hd.
I have a dedicated 2mb link and when i get over 8-10 simultaneous calls the agents start complainig about breakage and low volume calls. I'm not using any kind of compression.
I tested bypassing the vicidial server, connecting directly to my SIP provider(passing through my nat) and started to get perfect call quality. I also checked my nat networkcard at that time and it didn't even passed 1.3 mb of data being transfered which pretty much made me discard carrier and sip provider problems.
I have on this same box a digium board connected to a E1 and when I dial through the E1 i can have over 20 simultaneous calls without a problem.
I only make calls through E1 when I'm not making calls through SIP. I guess that with the digium board installed, i could discard timing issues also.
Is there a particular config for using with sip i'm missing here?
Thanks in advance,
Mauricio
Posted:
Fri Apr 09, 2010 10:39 am
by mflorell
loadavg on the server when this happens?
Posted:
Fri Apr 09, 2010 10:55 am
by dimorry
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
4936 root 20 0 37088 15m 4612 S 10 0.4 39:08.62 asterisk
4763 mysql 20 0 521m 97m 5176 S 6 2.4 14:45.05 mysqld
4972 root 20 0 9816 6624 2384 S 4 0.2 9:05.48 AST_update.pl
4981 root 20 0 9516 6324 2304 S 1 0.2 1:28.31 AST_VDauto_dial
17469 root 20 0 9184 6012 2348 S 1 0.1 0:00.21 AST_vm_update.p
8130 www-data 20 0 71140 8584 4948 S 1 0.2 0:03.61 apache2
9909 www-data 20 0 70848 6988 3620 S 1 0.2 0:03.00 apache2
10413 www-data 20 0 70848 7160 3772 S 1 0.2 0:03.18 apache2
17071 www-data 20 0 70848 6580 3240 S 1 0.2 0:00.34 apache2
19611 www-data 20 0 70848 7540 4040 S 1 0.2 0:11.67 apache2
4867 www-data 20 0 70848 7072 3704 S 0 0.2 0:04.92 apache2
9299 www-data 20 0 70844 7460 3996 S 0 0.2 0:03.59 apache2
11216 www-data 20 0 71152 8404 4760 S 0 0.2 0:15.99 apache2
13094 www-data 20 0 70848 6996 3640 S 0 0.2 0:01.92 apache2
13958 www-data 20 0 70844 6976 3612 S 0 0.2 0:01.44 apache2
14007 www-data 20 0 70848 7600 4088 S 0 0.2 0:14.17 apache2
14587 www-data 20 0 71148 8388 4748 S 0 0.2 0:13.60 apache2
14651 www-data 20 0 70848 7540 4032 S 0 0.2 0:12.97 apache2
14705 www-data 20 0 71140 8340 4708 S 0 0.2 0:13.82 apache2
17077 www-data 20 0 70848 6192 2880 S 0 0.1 0:00.32 apache2
17484 root 20 0 2308 1136 856 R 0 0.0 0:00.01 top
22184 www-data 20 0 70848 7524 4032 S 0 0.2 0:10.86 apache2
29711 www-data 20 0 71148 8636 4992 S 0 0.2 0:08.10 apache2
30104 www-data 20 0 71112 7972 4368 S 0 0.2 0:07.91 apache2
30105 www-data 20 0 70848 7128 3736 S 0 0.2 0:08.07 apache2
30220 www-data 20 0 70848 7412 4012 S 0 0.2 0:08.26 apache2
Posted:
Thu Apr 22, 2010 9:49 am
by dimorry
I don't know a lot about linux, the information i posted is a print of the top command. Does this give you what you need?
Is there anything else I can post?
Thanks again.
Posted:
Thu Apr 22, 2010 11:03 am
by DarknessBBB
To ensure that you're using the Sangoma as timing source type this in your asterisk CLI:
- Code: Select all
dahdi show status
or, if you have Zaptel:
- Code: Select all
zap show status
I really have the same problem as yours, so I can tell all the tests that I've made
Posted:
Thu Apr 22, 2010 11:08 am
by dimorry
I have a digium board installed.
here is the print from zap show status
Description Alarms IRQ bpviol CRC4
T2XXP (PCI) Card 0 Span 1 OK 0 0 0
T2XXP (PCI) Card 0 Span 2 UNCONFIGUR 0 0 0
Posted:
Thu Apr 22, 2010 11:46 am
by DarknessBBB
dimorry wrote:I have a digium board installed.
here is the print from zap show status
Description Alarms IRQ bpviol CRC4
T2XXP (PCI) Card 0 Span 1 OK 0 0 0
T2XXP (PCI) Card 0 Span 2 UNCONFIGUR 0 0 0
It seems OK, take a look at "my" thread, I think we have the same problem...
http://www.vicidial.org/VICIDIALforum/v ... hp?t=10595
Posted:
Thu Apr 22, 2010 2:30 pm
by dimorry
Thanks for the help. I've found a problem when i ran the mtr command.
Apparently i'm loosing up to 7% of the packets in my NAT, and over 10% with my provider on the other edge.
Host Loss% Snt Last Avg Best Wrst StDev
1. 10.1.1.3 7.0% 115 0.2 0.3 0.1 1.2 0.2
2. ???
3. atm1-0-0-663.bsa.gvt.net.br 0.0% 115 7.9 8.0 7.2 9.8 0.7
4. gvt-ge-0-0-0.rc02.bsa.gvt.net.br 0.0% 115 8.6 8.1 7.3 11.2 0.7
5. gvt-ge-4-0-0.rc01.bsa.gvt.net.br 0.0% 114 174.3 48.1 42.4 174.3 19.6
6. gvt-ge-0-1-0.rc01.cta.gvt.net.br 0.0% 114 41.5 42.1 41.5 43.6 0.4
7. gvt-ae-3-rd01.cta.gvt.net.br 0.0% 114 41.1 45.8 40.1 149.7 19.3
8. gvt-host.gvt.net.br 11.6% 114 42.5 106.6 40.4 597.6 139.9