I deleted an entry from clusterer table , but still opensips try to ping that node .

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

I deleted an entry from clusterer table , but still opensips try to ping that node .

Sasmita Panda
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

SamyGo
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

Sasmita Panda
Hi   Sammy,

Yes , you are right . I need to reload the cluster data through MI command . After reloading its seems fine . 

I was not aware about the fact that the cluster data also get shared with all nodes when I am adding that in 1 node only . 

Thank you for your explanation . Its really helpful . 


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


On Mon, Dec 31, 2018 at 10:25 PM SamyGo <[hidden email]> wrote:
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

Sasmita Panda
Hi, 
I have another doubt . Please do help me . 

When I am reading usrloc module document , its saying in a cluster if we want to replicate the contacts across the cluster then we have to set a parameter as below .
modparam("usrloc", "replicate_contacts_to", 1)
The default value is 0 , where no cluster id is mentioned . 
I have not set this , I have a cluster having 2 node . While I am registering a user , the contact is getting replicated between 2 nodes . 
If I am trying to mention this parameter , then opensips is not getting started . Its saying 
Parameter <accept_replicated_contacts> not found in module <usrloc>
So , my question is , if this parameter is not set , still how contact replication is happening ? Is this the default behavior of cluster module ? 
May be my question is foolish ,it will be great if anybody will explain this . 


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


On Wed, Jan 2, 2019 at 12:22 PM Sasmita Panda <[hidden email]> wrote:
Hi   Sammy,

Yes , you are right . I need to reload the cluster data through MI command . After reloading its seems fine . 

I was not aware about the fact that the cluster data also get shared with all nodes when I am adding that in 1 node only . 

Thank you for your explanation . Its really helpful . 


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


On Mon, Dec 31, 2018 at 10:25 PM SamyGo <[hidden email]> wrote:
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

Vlad Patrascu

Hi Sasmita,

By default, there is no clusterer replication if "replicate_contacts_to" parameter is not set in usrloc. Also, even if another node is sending replication packets, they will no get processed on the receiving node unless "accept_replicated_contacts" is set. On a typical setup, both these parameter should be set on all nodes.

Are you getting any other errors in the logs besides that "parameter not found" ?

Btw, I strongly suggest updating to 2.4 as it has received major upgrades in terms of clustering.

Regards,

Vlad Patrascu
OpenSIPS Developer
http://www.opensips-solutions.com
On 01/02/2019 12:10 PM, Sasmita Panda wrote:
Hi, 
I have another doubt . Please do help me . 

When I am reading usrloc module document , its saying in a cluster if we want to replicate the contacts across the cluster then we have to set a parameter as below .
modparam("usrloc", "replicate_contacts_to", 1)
The default value is 0 , where no cluster id is mentioned . 
I have not set this , I have a cluster having 2 node . While I am registering a user , the contact is getting replicated between 2 nodes . 
If I am trying to mention this parameter , then opensips is not getting started . Its saying 
Parameter <accept_replicated_contacts> not found in module <usrloc>
So , my question is , if this parameter is not set , still how contact replication is happening ? Is this the default behavior of cluster module ? 
May be my question is foolish ,it will be great if anybody will explain this . 


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


On Wed, Jan 2, 2019 at 12:22 PM Sasmita Panda <[hidden email]> wrote:
Hi   Sammy,

Yes , you are right . I need to reload the cluster data through MI command . After reloading its seems fine . 

I was not aware about the fact that the cluster data also get shared with all nodes when I am adding that in 1 node only . 

Thank you for your explanation . Its really helpful . 


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


On Mon, Dec 31, 2018 at 10:25 PM SamyGo <[hidden email]> wrote:
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

Mohit Sachan
hi 
Can you suggest me how to install oversip for webrtc for opensips-2.4 in centos.

On Thu, Jan 3, 2019 at 10:56 PM Vlad Patrascu <[hidden email]> wrote:

Hi Sasmita,

By default, there is no clusterer replication if "replicate_contacts_to" parameter is not set in usrloc. Also, even if another node is sending replication packets, they will no get processed on the receiving node unless "accept_replicated_contacts" is set. On a typical setup, both these parameter should be set on all nodes.

Are you getting any other errors in the logs besides that "parameter not found" ?

Btw, I strongly suggest updating to 2.4 as it has received major upgrades in terms of clustering.

Regards,

Vlad Patrascu
OpenSIPS Developer
http://www.opensips-solutions.com
On 01/02/2019 12:10 PM, Sasmita Panda wrote:
Hi, 
I have another doubt . Please do help me . 

When I am reading usrloc module document , its saying in a cluster if we want to replicate the contacts across the cluster then we have to set a parameter as below .
modparam("usrloc", "replicate_contacts_to", 1)
The default value is 0 , where no cluster id is mentioned . 
I have not set this , I have a cluster having 2 node . While I am registering a user , the contact is getting replicated between 2 nodes . 
If I am trying to mention this parameter , then opensips is not getting started . Its saying 
Parameter <accept_replicated_contacts> not found in module <usrloc>
So , my question is , if this parameter is not set , still how contact replication is happening ? Is this the default behavior of cluster module ? 
May be my question is foolish ,it will be great if anybody will explain this . 


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


On Wed, Jan 2, 2019 at 12:22 PM Sasmita Panda <[hidden email]> wrote:
Hi   Sammy,

Yes , you are right . I need to reload the cluster data through MI command . After reloading its seems fine . 

I was not aware about the fact that the cluster data also get shared with all nodes when I am adding that in 1 node only . 

Thank you for your explanation . Its really helpful . 


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


On Mon, Dec 31, 2018 at 10:25 PM SamyGo <[hidden email]> wrote:
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: I deleted an entry from clusterer table , but still opensips try to ping that node .

Sasmita Panda
HI , 

The data get stored in the cache for use of cluster module won't get deleted . 

Like , when I am saving something on location table that get deleted when the user get logout . 

The reference cfg file what I get for cluster module is only SET the data in the cache and GET that from cache . 

How can I remove the data from the cache ? 

Attached my config file . Please have a look into it . Let me know if I am doing anything wrong . 


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


On Thu, Jan 17, 2019 at 11:18 AM Mohit Sachan <[hidden email]> wrote:
hi 
Can you suggest me how to install oversip for webrtc for opensips-2.4 in centos.

On Thu, Jan 3, 2019 at 10:56 PM Vlad Patrascu <[hidden email]> wrote:

Hi Sasmita,

By default, there is no clusterer replication if "replicate_contacts_to" parameter is not set in usrloc. Also, even if another node is sending replication packets, they will no get processed on the receiving node unless "accept_replicated_contacts" is set. On a typical setup, both these parameter should be set on all nodes.

Are you getting any other errors in the logs besides that "parameter not found" ?

Btw, I strongly suggest updating to 2.4 as it has received major upgrades in terms of clustering.

Regards,

Vlad Patrascu
OpenSIPS Developer
http://www.opensips-solutions.com
On 01/02/2019 12:10 PM, Sasmita Panda wrote:
Hi, 
I have another doubt . Please do help me . 

When I am reading usrloc module document , its saying in a cluster if we want to replicate the contacts across the cluster then we have to set a parameter as below .
modparam("usrloc", "replicate_contacts_to", 1)
The default value is 0 , where no cluster id is mentioned . 
I have not set this , I have a cluster having 2 node . While I am registering a user , the contact is getting replicated between 2 nodes . 
If I am trying to mention this parameter , then opensips is not getting started . Its saying 
Parameter <accept_replicated_contacts> not found in module <usrloc>
So , my question is , if this parameter is not set , still how contact replication is happening ? Is this the default behavior of cluster module ? 
May be my question is foolish ,it will be great if anybody will explain this . 


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


On Wed, Jan 2, 2019 at 12:22 PM Sasmita Panda <[hidden email]> wrote:
Hi   Sammy,

Yes , you are right . I need to reload the cluster data through MI command . After reloading its seems fine . 

I was not aware about the fact that the cluster data also get shared with all nodes when I am adding that in 1 node only . 

Thank you for your explanation . Its really helpful . 


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


On Mon, Dec 31, 2018 at 10:25 PM SamyGo <[hidden email]> wrote:
Hi,
Did you restart OpenSIPS process on both node1, and 2 simultaneously ? The way I look at this is one of the two nodes kept the 3rd one in the memory and restarting both nodes one at a time resulted in both sharing their node structure and hence node3 stayed visible. 
I think possible way to remove a node gracefully would be to disable the node via the MI command and then remove from DB. I will try doing this on my test setup as well.

Regards,
Sammy 
 

On Fri, Dec 28, 2018 at 6:40 AM Sasmita Panda <[hidden email]> wrote:
Hi All, 

I have a cluster of 2 nodes . Both in working condition .    Then I added another node in the same cluster which is down . 

I restarted the opensips process , so it starts pinging the new node to check its status .  As the new node is down  , other nodes in the cluster wont get any reply for the ping  . Then I remove the 3rd node from the cluster table and restart the opensips process . 

Now what I am getting in logs is , still the 2 working node in the cluster try to ping the 3rd node which is not in the DB .  

Is this an issue on the cluster module or I am doing something wrong ?? Please help me . 
Thanks & Regards
Sasmita Panda
Senior Network Testing and Software Engineer
3CLogic , ph:07827611765
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users

opensips-cluster-new.cfg (29K) Download Attachment