locked
lync call forwarding and simring produces 15 sec lag RRS feed

  • Question

  • Hi
    I have a OCS 2007 R2 merged with Lync enterprise deployment with enterprise voice via sip trunking. On OCS everything works. On Lync incoming and outgoing calls works great. However forwarding and simultaneous ring produces a 15 sec lag.
    Example: a user has phone number 12345678 and uses simring to his mobile 98765432. When someone calls 12345678, either from lync or mobile, it starts to ring on the users pc right away but it takes 15 seconds before his mobile starts to ring.
    When i trace i can see the incoming INVITE to 12345678 on the mediationserver and front end. Half a second later i see the outgoing INVITE to 98765432 on the front end, but on the mediation server it shows up 12 seconds later! The same goes for forwarding a call.
    Anyone know what this can be caused by and how to fix it?

    Regards
    Jonatan

    Thursday, June 9, 2011 9:54 AM

Answers

  • So i was finaly able to sort out the problem with the help of microsoft.
    It was actually a 15 sec lag on all calls routed through lync, the reason i only noticed it on sim ring and so on was because we have two sip trunks and two phonenumber series and they where routed differently.
    Either way the problem was that the mediation servers was trying to connect to the edge internal interface. Mediation and edge are in different subnets and mediation didnt have a proper route to the edge internal interface so basicly the mediation server held all the calls for 10 seconds due to timeouts. The other 5 seconds is what it takes to connect the calls.

    Regards
    Jonatan

    • Marked as answer by Jonatan V Tuesday, July 19, 2011 7:05 AM
    Tuesday, July 19, 2011 7:05 AM

All replies

  • Hi,

    I suspect a misconfiguration of your normalization rules, or your routing.

    I can't see the same error in my own system. Can you tell, from tracing, what the FE server is doing with the call. Where it is trying to send it, initially?

    KR,

    Lasse Wedø


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. If the answer wasn't exactly what you were looking for, but gave you enough information to move on, please vote the post as "helpful".

    This can be beneficial to other community members reading the thread.
    Friday, June 10, 2011 7:37 AM
  • Hi Lasse
    Thanks for your reply.

    My voice policy looks like this
    Enabled calling features: call forwarding, delegation, call transfer, call park, simultaneous ring, team call, pstn reroute. Disabled: bandwith policy override, malicious call tracing.
    Associated PSTN usage pattern ^\+

    Dialplan looks like this
    Starting digits: "Nothing", Length: Any, Digits to remove: 0, Digits to add: +47, Pattern to match: ^(\d*)$, Translation rule: +47$1, Internal extension: No

    There are three INVITEs on the FE, one outgoing starting with sip:+4798765432@domain.com:5061, one incoming with sip:+4798765432@domain.com:5061 and one outgoing with sip:+4798765432@"pstn gw ip":5070
    Wich one do i check, and whats the line where i can verify where its being sent?

    Regards
    Jonatan

    Friday, June 10, 2011 8:07 AM
  • Hi,Jonatan,

    Did you enable media by pass?

    Please check your voice route,trunk configuration setting with the guide of Direct_SIP document in the following download link(Especially the page 11-18)

    http://www.microsoft.com/downloads/en/details.aspx?FamilyID=8c64a1e1-f0b3-479c-a265-e480875c61d8&displaylang=en

    Regards,

    Sharon


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.
    Tuesday, June 14, 2011 3:21 AM
  • Hi Sharon

    We dont use direct sip, we use sip trunking from one of the supported providers. In our end we have a mediation pool consisting of two mediation servers.
    And no, media bypass is disabled. Same with call admission control.

    Regards
    Jonatan

    Tuesday, June 14, 2011 11:58 AM
  • Hi Jonatan,

    Does this number: "+4798765432" happen to be a number belonging to a user in your system?

    The call should have been sent directly to the pstn. In my "book" You should only have seen the +4798765432@"pstn gw ip":5070 call. Not the internal ones.

    One final check; What's the client's setting on cfwd and simring. It is possible to adjust how long the system waits until sending the call out.

     

    KR,

    Lasse Wedø


    Please take a second to hit the green arrow on the left if the post was helpful, or mark it as an answer if it resolved your issue.

    This can be beneficial to other community members reading the thread.
    Tuesday, June 14, 2011 6:29 PM
  • Hi Lasse

    That number represents a users mobile phone number. So no its not present in my system, with the exception of on the users phone tab in AD.

    Cant say i know if it should be this or that, but i see the same behavior when i trace a user on OCS (the three invites). The only difference is that in OCS its the mediation server fqdn and not the pstn gw ip. And its no delay.

    For the example i have only activated simring in the client. But i see the same delay if i try to forward or transfer a call to the pstn on Lync.

    BTW skal hilse fra Geir Stenersen

    Regards
    Jonatan

    Wednesday, June 15, 2011 8:00 AM
  • Hi again Jonatan (And Geir ;)

    Sorry for my absence, I have been on a short holiday.

    If I get your configuration correct; You have a side by side implementation running at the moment. Have you tried to trace the call forward from Lync on other servers? I am just wondering if it tries the OCS 2007 environment first, and is waiting for some timeout first.

    Alternatively, would it be possible to shut down ocs 2007 services all together, and then try the same call? Just to rule out some 2007 involvement?

    I recall having seen this in a previous lab environment, but the problem went away after a completed migration (and removal of OCS 2007).

    I also want to know if this issue is related to only migrated users, or if the problem exists on a purely Lync enabled user (created after Lync was implemented, and enabled only through CSCP?)

    I know these questions might seem a bit far fetched, but I have not been able to reproduce the issue myself, and I have not been able to find other articles related to the issue.

    KR,

    Lasse Wedø


    Please take a second to hit the green arrow on the left if the post was helpful, or mark it as an answer if it resolved your issue.

    This can be beneficial to other community members reading the thread.
    Monday, June 20, 2011 7:06 PM
  • Hi again

    No problem, hope it was a good one.
    Yes, side by side ocs 2007 r2 and lync configuration. Ill do a trace tomorrow and let you know my findings.
    To me it looks like the delay is happening on the lync mediation server, and might very well be caused by timeouts. But what/why i dont know.
    A complete shutdown of ocs would be difficult, but ill look into it.
    Regarding users, the problem applies to both migrated and newly created users thats enabled for lync (cscp).
    I havent found anything useful myself, so i started a supportcase against ms today.

    Regards
    Jonatan


    UPDATE:
    Just did a trace on the OCS servers, but i cant say i can see anything interesting..
    Monday, June 20, 2011 8:39 PM
  • Too bad,

    Then let's hope you get something out of your support case ;)

    KR,

    Lasse Wedø


    Please take a second to hit the green arrow on the left if the post was helpful, or mark it as an answer if it resolved your issue.

    This can be beneficial to other community members reading the thread.
    Tuesday, June 21, 2011 7:22 PM
  • On a second thought; Have run the Best Practice analyzer? To verify all of your configuration?

    http://technet.microsoft.com/en-us/library/gg558584.aspx

    KR,

    Lasse Wedø


    Please take a second to hit the green arrow on the left if the post was helpful, or mark it as an answer if it resolved your issue.

    This can be beneficial to other community members reading the thread.
    Tuesday, June 21, 2011 7:42 PM
  • Yes i have. Nothing to worry about as far as i know. I have three warnings: "Lync 2010 global AD settings are stored in the system container. this configuration is for legacy users only",  "replication did not complete on 'grouopchatserver'" and "pool fully qualified domain name (fqdn) 'lyncpool.domain.com" is not found as a http service principal name (spn) on any user or computer. kerberos web authentication is not configured.."
    I hope ms gets back to today, and ill keep this post updated with the progress.

    Regards
    Jonatan

    Wednesday, June 22, 2011 8:01 AM
  • So i was finaly able to sort out the problem with the help of microsoft.
    It was actually a 15 sec lag on all calls routed through lync, the reason i only noticed it on sim ring and so on was because we have two sip trunks and two phonenumber series and they where routed differently.
    Either way the problem was that the mediation servers was trying to connect to the edge internal interface. Mediation and edge are in different subnets and mediation didnt have a proper route to the edge internal interface so basicly the mediation server held all the calls for 10 seconds due to timeouts. The other 5 seconds is what it takes to connect the calls.

    Regards
    Jonatan

    • Marked as answer by Jonatan V Tuesday, July 19, 2011 7:05 AM
    Tuesday, July 19, 2011 7:05 AM
  • Hi Jonatan,

    Good to hear you figured things out. I am just a bit puzzeled whe we did not see this in the debugs :)

     


    Lasse Wedø,
    Blog:Tech@work, Twitter: @lawedo

    Please take a second to hit the green arrow on the left if the post was helpful, or mark it as an answer if it resolved your issue.
    Wednesday, July 20, 2011 10:47 AM