[OpenSIPS-Users] Load balancer probe _mode=1 bug?
Bogdan-Andrei Iancu
bogdan at voice-system.ro
Mon Nov 8 13:24:34 CET 2010
Hi Bill,
What is the probing interval you configured ?
http://www.opensips.org/html/docs/modules/1.6.x/load_balancer.html#id250040
Could you test the attached patch ? (same as revision 7356 on trunk).
Regards,
Bogdan
Bill W. wrote:
> Hello Bogdan,
>
> This is how I mark the destination down:
>
> failure_route[1] {
> if (t_check_status("(408)|(5[0-9][0-9])")) {
> lb_disable();
> if (load_balance("1","sip","1")) {
> t_on_failure("1");
> t_relay();
> } else {
> t_reply("503", "Service Unavailable");
> }
> }
> }
>
>
> opensipsctl fifo lb_list shows: enabled=no auto-re=on
>
> I do not see any probes from opensips when probe_mode in the database
> is set to "1" which should probe the destination when it's marked down.
> When probe_mode is set to "2", and opensips probes continuously,
> it works fine.
>
> The docs say: "(1) - probing only when the destination is in disabled
> mode (disabling via MI command will competely stop the probing also).
> The destination will be automatically re-enabled when the probing will
> succeed next time;"
>
> So when I disable via lb_disable(), it is acting like I did it from the
> MI, and disabling the probing too.
>
> Thoughts?
>
> Thanks,
> Bill
>
> On 8/1/10 1:14 PM, Bogdan-Andrei Iancu wrote:
>
>> Hi Bill,
>>
>> How do you mark down the destination (MI ?) and what state are you setting ?
>>
>> Also, after marking it down, do you see any new pings being sent to that
>> destination?
>>
>> Something was changed during rev 6653:
>> http://opensips.svn.sourceforge.net/opensips/?rev=6653&view=rev
>>
>> But the new code seams ok for me - the LB_DST_STAT_NOEN_FLAG flag is set
>> when disabling the destination via MI and with this flag set, no more
>> pings are sent.
>>
>> Regards,
>> Bogdan
>>
>> Bill W wrote:
>>
>>> Hello Bogdan,
>>>
>>> I'm experiencing the same problem as I did before, with the load
>>> balancer not re-enabling a device marked down.
>>>
>>> If I go into the database and set probe_mode = 2, it re-enables the device.
>>>
>>> I'm currently running opensips 1.6.2-tls (x86_64/linux)
>>>
>>> Did the patch get reverted?
>>>
>>> Thanks!
>>> Bill
>>>
>>> Bogdan-Andrei Iancu wrote:
>>>
>>>
>>>> Super! I uploaded the fix on SVN.
>>>>
>>>> Thanks and regards,
>>>> Bogdan
>>>>
>>>> Bill W wrote:
>>>>
>>>>
>>>>> Hey Bogdan,
>>>>>
>>>>> It looks like that fixed it. Thanks so much!
>>>>>
>>>>> Bill
>>>>>
>>>>>
>>>>> Bogdan-Andrei Iancu wrote:
>>>>>
>>>>>
>>>>>
>>>>>> Hi Bill,
>>>>>>
>>>>>> Could you please try the attached patch? It seams that there was an
>>>>>> issue with the the probing values in the code. Let me know if the patch
>>>>>> does solves your problem and I will upload it on SVN.
>>>>>>
>>>>>> Regards,
>>>>>> Bogdan
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> Users mailing list
>>>>> Users at lists.opensips.org
>>>>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users at lists.opensips.org
>>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>
>>>
>>>
>>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
>
--
Bogdan-Andrei Iancu
OpenSIPS Bootcamp
15 - 19 November 2010, Edison, New Jersey, USA
www.voice-system.ro
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lb.patch
Type: text/x-diff
Size: 545 bytes
Desc: not available
Url : http://lists.opensips.org/pipermail/users/attachments/20101108/f2ad5290/attachment.patch
More information about the Users
mailing list