#443 Snom re-registering with Multiple Location Entries

trunk
closed-fixed
None
5
2011-11-04
2011-11-02
No

In the past Snom phones worked fine when it came to Registering with OpenSIPS. Now it appears no matter what version of code Snom is running it will Register and then you will have multiple Registers in the Location table. So if you set the Register time to 600 the first time it registers everything will be good. Then at 300 seconds the Snom phone sends a re-register and that one is added to the Location table, but the old register is still in the Location table also. So at 300 seconds you have 2 Location Entries for one phone. Then once you get to around 600 seconds a third re-register is sent and for a short while you have 3 Location entries for a single phone. I notice that Blink and Jitsi don't have this issue. I looked at the SIP REGISTER messages for Jitsi, Blink and Snom and I don't see any major difference that could cause this.

Please refer to the following mailing list to see the thread for this issue.
http://opensips-open-sip-server.1449251.n2.nabble.com/Multiple-Registered-Locations-for-a-single-agent-td6950474.html

On my main.c file it says that the version is 8267

Discussion

  • Vladut-Stefan Paiu

    • assigned_to: nobody --> vladut-paiu
     
  • Vladut-Stefan Paiu

    Hello,

    I was thinking that maybe your issue has something to do with the GRUU additions on the Registrar & Usrloc modules,
    but 8267 is before adding the GRUU support.

    Are you, by any change, using the query insert buffering support ? Also, what usrloc DB mode are you using ?

    When the DB records are broken, if you do opensipsctl ul show , do you still get the duplicate contacts there as well ? What I'm trying to see if it's just the DB info that is bogus, or if the OpenSIPS internal memory is affected as well.

    Regards,
    Vlad

     
  • Duane Larson

    Duane Larson - 2011-11-03

    I am using the following DB mode
    modparam("usrloc", "db_mode", 3)

    As for the Query Insert buffering support I am guessing you are talking about this
    http://www.opensips.org/Resources/InsertBuffering
    No I am not using that. Wasn't aware of it but it sounds interesting.

    I wouldn't say "When the DB records are broken" because it seems like it is a given that two location records are going to be in the MySQL Location table after half the re-register time has passed. If I do a "opensipsctl ul show" I don't see anything since I am not putting anything in memory

    Proxy02:~# opensipsctl ul show
    Domain:: location table=512 records=0

     
  • Vladut-Stefan Paiu

    Hello,

    Is it possible that you update to the latest OpenSIPS revision from trunk ? And also, please use the latest DB schema for usrloc.

    Why I'm asking you to do this is because there have been some bug-fixes related to the usrloc schema and also with the usrloc while in DB mode 3 and the issue might be fixed in the latest trunk.

    Please do this and let me know if the issue still appears.
    Regards,
    Vlad

     
  • Duane Larson

    Duane Larson - 2011-11-03

    Just downloaded Revision 8527.

    I restarted OpenSIPS now. Here is what my MySQL "verison" table says

    mysql> select * from version;
    +-----------------+---------------+
    | table_name | table_version |
    +-----------------+---------------+
    | watchers | 4 |
    | userblacklist | 2 |
    | dialog | 7 |
    | subscriber | 7 |
    | active_watchers | 10 |
    | presentity | 5 |
    | rls_watchers | 2 |
    | b2b_entities | 1 |
    | pua | 8 |
    | carrierroute | 3 |
    | load_balancer | 2 |
    | domainpolicy | 3 |
    | missed_calls | 4 |
    | usr_preferences | 3 |
    | route_tree | 2 |
    | acc | 6 |
    | dispatcher | 4 |
    | address | 5 |
    | pdt | 2 |
    | speed_dial | 3 |
    | globalblacklist | 2 |
    | rls_presentity | 1 |
    | lcr | 3 |
    | location | 1007 |
    | silo | 6 |
    | dr_groups | 2 |
    | dr_gw_lists | 1 |
    | b2b_logic | 1 |
    | sip_trace | 3 |
    | grp | 3 |
    | xcap | 4 |
    | domain | 2 |
    +-----------------+---------------+

     
  • Duane Larson

    Duane Larson - 2011-11-03

    Hmmm I deleted my old location table in MySQL and added in the updated version and now no phone can register. I see that when they send the Register message with the authentication stuff they still get a 401 Unauthorized. The weird thing is that even though they got a 401 they do show up in the location table. Not sure what the hell I did.

     
  • Duane Larson

    Duane Larson - 2011-11-03

    This new issue has something to do with my memcache password storage for registers. Not sure why dumping the old location table and updating it would have messed this up. Will have to figure this out. Then we will see if all this made a difference.

     
  • Duane Larson

    Duane Larson - 2011-11-03

    OK. I have no freakin clue what happened with memcache. seems like things weren't working and now everything appears to be fine. Weird. We'll see now if things are good.

     
  • Duane Larson

    Duane Larson - 2011-11-03

    I hate to say it but I think upgrading has broken my memcache setup for Register Passwords and Invite Passwords. When a Register goes through my script the xlog output is always <null>. I can't see that updating the Location database had anything to do with this so I think it might be the SVN code.

     
  • Duane Larson

    Duane Larson - 2011-11-04

    Vlad,

    Good news. I guess when I dropped my old Location table and then created the new table with the newer schema it fixed the problem. I don't have duplicates with my Snom phones anymore. Please remember that I reverted back to my original OpenSIPS Trunk version of 8267. So I am not on the latest. The lastest version has an issue with Memcache unless I've done something else stupid.

    The last 3 years that I have been building my OpenSIPS config for all the stuff I want it to do I haven't really thought to update any of the table schema's except for the "version" table or any other table that barked at me when I upgraded and tried to start OpenSIPS. I guess I need to go back and look at all my table schemas.

    So what should I do about the memcache issue? Should I open another ticket?

    Thanks for the help with this.

     
  • Vladut-Stefan Paiu

    Hello Duane,

    I'm glad that updating the table has fixed your issue. There was a problem with the keys for the DB table, so glad that it's ok now.

    After rev r8434 , OpenSIPS has been enhanced with a new Key-Value interface, so the memcached module has been removed and replace by cachedb_memcached. Also the code has significantly changed in the core. It might have been a problem were you kept using the old memcached module, but not sure. So please open a new ticket for your memcached issue, where you more accurately describe the issue you have.

    I will close the current bug report.

    Regards,
    Vlad

     
  • Vladut-Stefan Paiu

    • status: open --> closed-fixed
     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks