Resyncing the jb. last_delay -19, this delay -1503, threshol

[scenarist]

I use IAX extension connected to static IP (DEDICATED SERVER)
Softphone: Zoiper
Internet speed: download 100 Mbps | Upload 100 Mbps
Account entry:

Code: Select all

[xxxxxx]
defaultuser=xxx
fromuser=xxxx     
secret=xxxx
type=peer
dtmfmode=rfc2833
insecure=port,invite
host=xxxxx
defaultip=xxxx
port=5060
session-timers=refuse
disallow=all
allow=alaw
allow=ulaw
alow=gsm


I have a big problem with "breaking up" call in the middle of call.
I am receiving 5-8 complaints a day of calls "breaking up" from agents.
Asterisk CLI shows me the following:

Code: Select all

Resyncing the jb. last_delay -19, this delay -1503, threshold 1006, new offset 1554

[0K[Apr 29 13:22:32] WARNING[8856]: chan_iax2.c:906 jb_warning_output: Resyncing the jb. last_delay 11, this delay 1147, threshold 1056, new offset 407

[Apr 29 13:23:31] WARNING[8851]: chan_iax2.c:906 jb_warning_output: Resyncing the jb. last_delay 647, this delay 4282, threshold 2302, new offset -4282

[Apr 29 13:26:02] WARNING[8859]: chan_iax2.c:906 jb_warning_output: Resyncing the jb. last_delay -2, this delay -1762, threshold 1058, new offset 2169


in my iax.conf I have the following, maybe I must change something

Code: Select all

jitterbuffer=yes
forcejitterbuffer=no
maxjitterbuffer=1000
maxjitterinterps=10
resyncthreshold=1000

I don't know what could be the problem ? any suggestion ?

[williamconley]

check your cpu usage and your bandwidth and your i/o buffer. something is likely overloading. worthy of note is that most systems usually only hold 25 agents, you may merely need to move the DB to another server or add another dialer to your server to start a vicidial cluster.

of note: 1TB drives are usually quite slow. bigger is not better when it compromises the throughput speed of the data.

good job posting your specs!

[scenarist]

I found this on forum http://support.sharedband.com/kb/kb6/Poor+call+quality+on+VoIP


You may get problems that audio sounds broken, this is generally caused by out of sequence packets.

With most systems this can be fixed by using either of these methods:

a) Enable a Jitter Buffer

With FreePBX (Asterisk based Distro's like Trixbox, Elastix etc) you can change this in the sip_general_custom.conf

here are the variables you will require:

;Enable Jitter Settings
jitterbuffers=4
jbenable=yes
jbforce=yes
jbimpl=fixed
;jbimpl=adaptive
jbmaxsize=50
jbresyncthreshold=100
jblog=yes
Asterisk sip jbenable = yes|no
Enables the use of a jitterbuffer on the receiving side of a SIP channel. (Added in Version 1.4)

Asterisk sip jbforce = yes|no
Forces the use of a jitterbuffer on the receive side of a SIP channel. Defaults to "no". (Added in Version 1.4)

Asterisk sip jbmaxsize = Number
Max length of the jitterbuffer in milliseconds. (Added in Version 1.4)

Asterisk sip jbresyncthreshold = Number
Jump in the frame timestamps over which the jitterbuffer is resynchronized. Useful to improve the quality of the voice, with big jumps in/broken timestamps, usually sent from exotic devices and programs. Defaults to 1000. (Added in Version 1.4)

Asterisk sip jbimpl = fixed|adaptive
Jitterbuffer implementation, used on the receiving side of a SIP channel. Two implementations are currently available - "fixed" (with size always equals to jbmaxsize) and "adaptive" (with variable size). Defaults to fixed. (Added in Version 1.4)

Asterisk sip jblog = no|yes
Enables jitterbuffer frame logging. Defaults to "no". (Added in Version 1.4)

b) Utilise G729 Codec

The codec that you are using may be intolerant (G711 being a known example) of the disordered packets that are inevitable when using Sharedband. Please arrange for a more tolerant codec to be employed within your VOIP system. A known example of a more tolerant codec is G729.

If you have any further problems please Raise A Ticket


only one time my load average was 100 % and after restart server everything is ok. Ram used max 3GB of 8GB. Internet bandwidth 100/100Mbps on server.
I will try to use SIP extension instead of IAX and we will see....
@Williamconley
How much agents can I have on one single server with performances
Intel® Core™ i7-920 Quadcore
incl. Hyper-Threading Technology
RAM 24 GB DDR3 RAMHard disks2 x 750 GB SATA 3 Gb/s HDD
(Software-RAID 1)
[manual dial, no recording,no form, only outbound calling]

[williamconley]

normally 25 agents per dialer.

however, with manual dialing you'll be able to get more. possibly many more.

but: turn off the "software RAID" it is pointless and counterproductive. Mirror RAID on a motherboard is even not really recommended, but software raid will degrade performance horribly. Get a real RAID controller or just turn it off and back up your system regularly.

do not store recordings on the vicidial server. use the ftp scripts provided to push all recordings to another storage server instead.

your vicidial server is expected to have Small Fast drives, not huge (slow) storage drives.

and turn off hyperthreading. that's great for fooling Windows into thinking it has two cores for each physical core, but linux servers that use the entire processor can be insulted at being lied to ... and when they try to use that phantom processor they can demonstrate that insult. LOL (Plus, I've never seen an increase in capacity by hyperthreading ...).