Hi Everyone,
Today I felt a strange problem with my OpenSIPS load balancer that it started
giving me 503 Service Unavailable error and rejecting all the calls, & I
couldn't see anything in logs. I have not restarted it yet but just want to
know the actual reason of it if anybody has faced the same problem before.
One more thing that in the MySQL load_balancer table, I have given resource
pstn=100 for each destination server but the calls are more than this so would
that be effecting the load balancer & in a result failing all the calls. Should
I have to increase the resources for each server in load_balancer table??
Regards,
Faisal Rehman
________________________________
From: Faisal Rehman <[email protected]>
To: Sammy Govind <[email protected]>
Cc: OpenSIPS users mailling list <[email protected]>
Sent: Friday, January 6, 2012 3:05 PM
Subject: Re: [OpenSIPS-Users] Calls are not properly distributed using Load
Balancer
Hi Sammy,
I will definitely explore it and will share the conclusions with you :)
Regards,
Faisal Rehman
________________________________
From: Sammy Govind <[email protected]>
To: Faisal Rehman <[email protected]>
Cc: OpenSIPS users mailling list <[email protected]>
Sent: Friday, January 6, 2012 12:24 AM
Subject: Re: [OpenSIPS-Users] Calls are not properly distributed using Load
Balancer
Feels good to know that it wasnt really a big issue at all.
Regarding packet losses monitoring in opensips !! I wonder if you can use
siptrace module for that !! and guess I found this
http://lists.opensips.org/pipermail/news/2011-December/000165.html
on main page of opensips.org !!
Now I'm thinking if you can use this module, enable siptrace for all incoming
call and store+fwd that siptrace of the call to analyze, maybe you'll be able
to tell the packet losses in SIP communication.
On Thu, Jan 5, 2012 at 5:59 PM, Faisal Rehman <[email protected]> wrote:
Hi Sammy,
>
>
>The mistake was very minor as in the load_balancer table I was giving group id
>as 1 to all the destination servers while in dispatcher table the group id was
>same for the three servers as 1 but different for the two servers as 0 & 2, so
>that was the reason there were no calls on these two servers & was fixed
>assigning the same group id as 1 to all the servers & the calls were
>distributed equally after the second load test :)
>
>
>Ntop is I think for network monitoring & I installed it on one of my servers
>but don't know the exact mechanism of its monitoring for SIP especially. I am
>only dealing here with SIP yet so there isn't any built in tool or command in
>OpenSIPS for packet loss info.
>
>
>
>Regards,
>
>
>
>Faisal Rehman
>
>
>________________________________
> From: Sammy Govind <[email protected]>
>To: Faisal Rehman <[email protected]>; OpenSIPS users mailling list
><[email protected]>
>Sent: Thursday, January 5, 2012 5:29 PM
>
>Subject: Re: [OpenSIPS-Users] Calls are not properly distributed using Load
>Balancer
>
>
>
>Hi,
>
>
>Really nice to see your issue is resolved, mind sharing the mistake you had in
>table !?
>
>
> About the packet loss, try using some traffic monitoring tool for this like
>ntop, or some snmp based interface monitoring utility to find packet loss. If
>you are talking about call Media packet losses then you can get the stats if
>you're using RTPproxy. Rtpproxy publishes media stats at the end of the call.
>
>
>Regards,
>Sammy.
>
>
>On Thu, Jan 5, 2012 at 5:20 PM, Faisal Rehman <[email protected]>
>wrote:
>
>Hi Guys,
>>
>>
>>The issue is fixed & it was due to my mistake because I was using some
>>different values in the dispatcher table. By the way I have got one more
>>question which is that how can I check the total packet loss in OpenSISP for
>>the whole traffic?
>>
>>
>>
>>Regards,
>>
>>
>>Faisal Rehman
>>
>>
>>________________________________
>> From: Sammy Govind <[email protected]>
>>To: OpenSIPS users mailling list <[email protected]>
>>Sent: Thursday, January 5, 2012 11:25 AM
>>Subject: Re: [OpenSIPS-Users] Calls are not properly distributed using Load
>>Balancer
>>
>>
>>
>>Hi Faisal,
>>
>>
>>Please share the contents of load_balancer table. I doubt that the group
>>defined (or as Vlad doubt resource assigned)for the servers are different
>>thats why they are not receiving the traffic.
>>
>>
>>
>>Regards,
>>Sammy.
>>
>>
>>
>>On Wed, Jan 4, 2012 at 4:18 PM, Vlad Paiu <[email protected]> wrote:
>>
>>
>>>Hello,
>>>
>>>If using the load_balancer module, do the two servers that get no
traffic at all have the same assigned resources as the other 3
server that get even distribution of calls ?
>>>
>>>Regards,
>>>
>>>Vlad Paiu
OpenSIPS Developer
>>>
>>>On 01/04/2012 11:18 AM, Faisal Rehman wrote:
>>>Hi Guys,
>>>>
>>>>
>>>>I was testing my OpenSIPS loadbalancer in live production environment where
>>>>I was having 5 destination servers which should be getting equal calls load
>>>>but I was surprised to see that there were two servers where there was no
>>>>call but rest of the three servers were having equal traffic, so what can
>>>>be the possible reasons for that issue?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>Regards,
>>>>
>>>>
>>>>Faisal Rehman
>>>>
_______________________________________________
Users mailing list [email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>_______________________________________________
>>>Users mailing list
>>>[email protected]
>>>http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>
>>>
>>
>>_______________________________________________
>>Users mailing list
>>[email protected]
>>http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>
>>
>>
>>_______________________________________________
>>Users mailing list
>>[email protected]
>>http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>
>>
>
>
>
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
