[OpenSIPS-Users] Opensips -1.11 crash with mongodb replica set .

Sasmita Panda spanda at 3clogic.com
Wed Jun 8 09:25:16 CEST 2016 

Hi Liviu ,

      What are you suggesting for the mongodb config server ? Because confg
server is the one who contains the metadata of the shard cluster .

     Where should I put it ? Will I put this separately or in the opensips
box with mongos ? If I will put the config servers separately and each
mongos instance will get connected to the same set of config server then
will it be fruitful or not ?

*Thanks & Regards*
*Sasmita Panda*
*Network Testing and Software Engineer*
*3CLogic , ph:07827611765*

On Tue, Jun 7, 2016 at 5:05 PM, Sasmita Panda <spanda at 3clogic.com> wrote:

> Thank you so much . It will be helpful for me .
>
> *Thanks & Regards*
> *Sasmita Panda*
> *Network Testing and Software Engineer*
> *3CLogic , ph:07827611765*
>
> On Tue, Jun 7, 2016 at 4:41 PM, Liviu Chircu <liviu at opensips.org> wrote:
>
>> To avoid those kind of problems, I would recommend putting one "mongos"
>> instance for each OpenSIPS machine, on the same box. This way, not only do
>> you avoid problems with the mongos machine going down, but you'll speed up
>> query times as well, making your OpenSIPS workers more responsive!
>>
>> Liviu Chircu
>> OpenSIPS Developerhttp://www.opensips-solutions.com
>>
>> On 07.06.2016 13:20, Sasmita Panda wrote:
>>
>> Hi Liviu ,
>>
>>      I just have one question .  I have implemented opensips with mongodb
>> shard cluster . I have one replica set with three member . I have three
>> config server for each and mongos service running .
>>
>>      What I wanted to know is , the server in which my mongos service is
>> running , If that machine is get down what will happen ?
>> I need to monitor that machine so that it wont go down ? Is this show ?
>>
>> *Thanks & Regards*
>> *Sasmita Panda*
>> *Network Testing and Software Engineer*
>> *3CLogic , ph:07827611765*
>>
>> On Mon, Jun 6, 2016 at 4:41 PM, Liviu Chircu <liviu at opensips.org> wrote:
>>
>>> Thank you, Sasmita! Will reply here as soon as there is progress on the
>>> matter.
>>>
>>> Liviu Chircu
>>> OpenSIPS Developerhttp://www.opensips-solutions.com
>>>
>>> On 06.06.2016 13:55, Sasmita Panda wrote:
>>>
>>> Hi ,
>>>
>>>  I have created an issue with detailed steps . Bellow is the issue ID .
>>> Let me know if the issue will get fixed .
>>>
>>>   https://github.com/OpenSIPS/opensips/issues/895
>>>
>>> Thanks
>>>
>>> *Thanks & Regards*
>>> *Sasmita Panda*
>>> *Network Testing and Software Engineer*
>>> *3CLogic , ph:07827611765*
>>>
>>> On Mon, Jun 6, 2016 at 3:32 PM, Sasmita Panda < <spanda at 3clogic.com>
>>> spanda at 3clogic.com> wrote:
>>>
>>>> Thank you so much for the information . Let me try to deploy with
>>>> mongos server .
>>>>
>>>> I will post the detailed steps for opensips crash as soon as I will be
>>>> free .
>>>>
>>>> *Thanks & Regards*
>>>> *Sasmita Panda*
>>>> *Network Testing and Software Engineer*
>>>> *3CLogic , ph:07827611765*
>>>>
>>>> On Mon, Jun 6, 2016 at 2:14 PM, Liviu Chircu < <liviu at opensips.org>
>>>> liviu at opensips.org> wrote:
>>>>
>>>>> Hi Sasmita!
>>>>>
>>>>> We have some plans to do some more testing on the module, including
>>>>> your "connect directly to replica set" usage case, and  also bringing it up
>>>>> to date with the latest mongo driver. However, it is hard to give you an
>>>>> estimation for when this work will start.
>>>>>
>>>>> Currently, the fastest way to fix your problems is for you to set up a
>>>>> "mongos" instance, along with 3 config server instances (these are needed
>>>>> by mongos, and ensure proper write consistency), and configure
>>>>> "cachedb_mongodb" to use this new "mongos" node.
>>>>>
>>>>> A mongos instance is aware of all the replica sets (aka "shards"). In
>>>>> your case, it will initially shard the keys to your single replica set -
>>>>> thus, it will hold 100% of the data. As your data set grows, you may
>>>>> partition it over to additional replica sets, which you can configure into
>>>>> mongos without modifying anything on the OpenSIPS side. This way, failover
>>>>> within a replica set will also happen transparently, again, without
>>>>> impacting OpenSIPS at all.
>>>>>
>>>>> PS: if you have a way to 100% reproduce a crash, please open a GH
>>>>> ticket describing the steps, and we'll go from there! [1]
>>>>>
>>>>> [1]: https://github.com/OpenSIPS/opensips/issues
>>>>>
>>>>> Liviu Chircu
>>>>> OpenSIPS Developerhttp://www.opensips-solutions.com
>>>>>
>>>>> On 06.06.2016 11:23, Sasmita Panda wrote:
>>>>>
>>>>> Hi All ,
>>>>>
>>>>>        I am using opesips-1.11 with mongodb replica set . I have 3
>>>>> members in the replica set , promary , secondary and arbitrary .
>>>>>
>>>>>       Problem 1 :  Sometime , If primay is not rechable , the
>>>>> secondary becomes primary , But opensips loss connection from mongodb . Its
>>>>> wont put any data in the data base . My call goes on but their is not data
>>>>> in the mongodb database .
>>>>>
>>>>>         Problme 2: If the primay machine is down then secondary
>>>>> becomes primary within some millisecond time  , but  opensips crashes
>>>>> giving bellow error
>>>>>
>>>>>       ERROR:cachedb_mongodb:mongo_con_get: Failed to run query. Err =
>>>>> 6, 0 , 0
>>>>>  CRITICAL:core:receive_fd: EOF on 10
>>>>>  INFO:core:handle_sigs: child process 5278 exited by a signal 11
>>>>>  INFO:core:handle_sigs: core was generated
>>>>>  INFO:core:handle_sigs: terminating due to SIGCHLD
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:sig_usr: signal 15 received
>>>>>  INFO:core:cleanup: cleanup
>>>>>  NOTICE:db_cachedb:destroy: destroy module db_cachedb ...
>>>>>  NOTICE:cachedb_mongodb:destroy: destroy module cachedb_mongodb ...
>>>>>
>>>>>      If my secondary mongodb machine is not reachable then rather
>>>>> connection with db breaks for sometime but opensips wont crashes . But in
>>>>> case of Primary opensips crashes with above error . Is this an expected
>>>>> behavior or I am doing anything wrong .
>>>>>
>>>>>       Data loss can be bearable but application cant be . So please
>>>>> let me know whats the problem .
>>>>>
>>>>> *Thanks & Regards*
>>>>> *Sasmita Panda*
>>>>> *Network Testing and Software Engineer*
>>>>> *3CLogic , ph:07827611765*
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Users mailing listUsers at lists.opensips.orghttp://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 listUsers at lists.opensips.orghttp://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 listUsers at lists.opensips.orghttp://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
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20160608/905ca99a/attachment-0001.htm>

More information about the Users mailing list