CUCM 8.6 MeetMe Software CFB not 'rolling over'.

[JGB_GtmK_CJoN]

Hi,

CUCM: 8.6.2.22030-1

I have a PUB & 1 SUB and I'm having unexpected results when maxing out MeetMe resources.  There is 1 MRGL for the main site (all testing conducted here).  This MRGL consists of 4 MRGs: in sequence -> SoftwarePub, SoftwareSub, Hardware1, MOH.  The Software MRG's have Annunciator, CFB, MTP in them for each corresponding node.  In MeetMe System parameters - max is set to 10 and this is working fine (limits each MeetMe bridge to 10 participants).  The IP Voice Media Streaming App is active on both Pub & Sub - Call Count is set to 48 and is working fine (limits a total of 48 streams on each Software CFB).

The problem is that when there are 48 active MeetMe participants in different bridges on the priority one Software CFB, a new MeetMe bridge can not be established.  In other words, the second Software MRG in the MRGL will not kick in and allow additional MeetMe bridges to be opened. 

I've changed the priority of the SoftwarePub & SoftwareSub to ensure that each node will handle their 48 streams - and they do.  During my testing, I started a MeetMe bridge with the  SoftwarePub in priority 1 spot, used RTMT to verify that the streams where being handled by the PUB, then without terminating the MeetMe bridge, moved the SoftwareSub up to priorpty 1 in the MRGL and opened a new MeetMe bridge and confirmed using RTMT that the new bridge was being handled by the Sub.

Based on research, with the current config, I should expect a total of 48 + 48 = 96 sessions/streams available for this cluster.  In theory, this can be sliced up by having 9 MeetMe bridges with 10 participants and a 10th MeetMe bridge with 6 psrticipants.

Please provide insight as to what I can do to ensure that both Software CFB's will handle bridges.

Thank you!!

[Manish Gogna · Accepted solution]

Hi,

Can you test by putting the SoftwareSub and SoftwarePub CFB in the same MRG and see if the same behavior is seen.

Manish

[JGB_GtmK_CJoN]

Hi Manish,

I'll test this out and let you know if it worked.

Thank you for the reply!

[JGB_GtmK_CJoN]

Hi Manish,  

This solved the issue.  I added a new MRG with only the SFT CFBs and set it in the top spot of the MRGL.  Now, MeetMe load balancing behavior is as expected.  Each soft conf bridge is taking a call in round robin fashion - i confirmed this via RTMT.  I expect that it is OK to leave the other MRGs in the MRGL as they were - pleas advise if you see any issues with me leaving them in.

Thank you very much Manish - I appreciate you assistance!!

[Manish Gogna]

Hi,

Thanks for the update. Yes, leaving the MRG's / MRGL's like this should be okay. Within the MRG the resources are allocated by the cucm by using an internal algorithm, once it is unable to find the resources in the top MRG it then looks into next MRG in the MRGL. In your case the latter part was not happening. If you wish you can let TAC investigate on why the cucm was unable to allocate resources from second MRG, else the current setup is absolutely fine.

Manish