locked
ASMCU Moderate / Maximum Capacity RRS feed

  • Question

  • Hello together,

    I currently encounter a strange problem in our deployment.

    We have 1 Pool with 3 FE Server.

    When I initiate a conference with an external client and share my screen, the appropriate FE Server who is hosting the AS conference, always shows me an event that the ASMCU is at moderate capacity. As soon as the connection is established I receive a new event which says, that the ASMCU ist a normal capacity. (20 seconds after the first event with the moderate capacity). When we have multiple clients connected (5 or 6) it also occured, that the ASMCU was at maximum capacity. In the reports I saw an internal ASMCU failure but I am not sure if this pointed to that maximum capacity event.

    I have checked the conferencing configuration, tried to change the maxbandwith parameter, but nothing solved the problem.
    My NIC performance on the FE is constantly under 3 MB/s so this can not be the problem. This problem occurs on all three Front End Server but only when someone from external connects to the ASMCU. Sharing between only internal users in a conference will not reproduce the problem...

    Did anyone of you noticed something like this before?

    Greets,
    Markus

    Wednesday, March 28, 2018 5:22 PM

Answers

  • Hey all,

    I would like to inform you that it seems like that I have just solved my problem with the ASMCU capacity.
    First of all, the KHI indicators are great.
    The only thing I mentioned there was the value: Packet Received Discarded:
    Our Servers are running in VMware and the following link helped me to reduce the amount of discarded packages.
    https://kb.vmware.com/s/article/2039495 --> Just for your information :)
    I have set up the values on one Server to max and it directly shows 0 discarded packages after a day. I will increase it gradually on the others too.

    Now back to the main problem.
    As I mentioned before this problem only occurs when s.o. from external establishes a connection to the ASMCU.
    I have just changed our Edge LB scenario to DNS LB. Before this change we had an unsupported HLB deployment. I have mentioned that the Media estbalishment with the ASMCU is now much faster than before and the Server did not show any kind of capacities failures after a lot of testing.

    So this seems to resolve my problem :) Thank you for your help!

    Greets and a nice weekend,

    Markus

    • Marked as answer by Markus_95 Saturday, April 7, 2018 2:33 PM
    Saturday, April 7, 2018 2:33 PM

All replies

  • It sounds like the hardware used for your Front End or Edge pool may not be up to standard. Review the following technet and compare it to your environment: https://docs.microsoft.com/en-us/skypeforbusiness/plan-your-deployment/requirements-for-your-environment/server-requirements. I would also recommend running KHI's for a few days and reviewing the results (https://blogs.technet.microsoft.com/praj/2015/04/24/lync-khi-performance-data-collection-and-analysis/). It would also be helpful to post what CU you are on.
    • Edited by Tyler Poston Thursday, March 29, 2018 1:30 AM
    • Proposed as answer by Alice-Wang Thursday, March 29, 2018 2:44 AM
    Thursday, March 29, 2018 1:30 AM
  • Hi,

    Agree with Tyler.

    Please check if you meet the requirement for skype for business server 2015.

    If so, please help us to confirm the following questions:
    1.When external user had SFB conference, what issue did you encounter?
    2.Are there any event IDs in SFB FE server?

    Another document for your reference :
    Capacity planning for Lync Server 2013 using the user models
    https://technet.microsoft.com/en-us/library/gg615015(v=ocs.15).aspx


    Best Regards,
    Alice Wang


    Please remember to mark the replies as answers if they helped. If you have feedback for TechNet Subscriber Support, contact tnsf@microsoft.com.


    Click here to learn more. Visit the dedicated forum to share, explore and talk to experts about Microsoft Teams.

    Thursday, March 29, 2018 2:57 AM
  • Hey you two,

    first of all, thanks for your answers!

    Our FE Server are virtualized in VMWare and currently running at 16 GB RAM.
    Constantly 70% of RAM is in use. I could set it up for testing because the recommended RAM is 32 GB.
    CPU is OK so far.

    Edge Server is currently at 50% RAM with 4 GB.

    I will configure and check the KHIs! :)

    We have installed the December CU from 2017. Unfortunately I can not install the March CU because we have some issues that will be resolved witht the Januar Hotfix CU and I have to receive the CU first from Microsoft.

    Our Edge Server Deployment is also not correct at the moment regarding the traffic flow, because we have a HLB where all the A/V traffic from the Edge goes through... - I will change this to DNS LB next week. Maybe this could also be one of the reasons.

    Regarding the issues for the users:
    In general I can not reproduce any issue. It is working. But I assume that when multiple user from external connect to the ASMCU at the same time, the maximum capacity will be reached and therefore it may comes to an internal ASMCU failure and the connection will be interrupted. But I am not pretty sure.

    The only Events I see on the Front End Server who owns the ASMCU for the conference are the Event IDs for the moderate/maximum capacity and after a few seconds for the "normal" capacity level. IDs (32031/32/33

    Greets,
    Markus




    • Edited by Markus_95 Thursday, March 29, 2018 11:02 AM removed picture
    Thursday, March 29, 2018 6:14 AM
  • Hi Markus_95,

    Thanks for your reply. According the current status, we suggest you could monitor for several days to see if it works now.

    I did many research, the scenario mostly may be related to the “MaxBandwidthPerAppSharingServiceMb”, and I noticed you have already changed it, you could try to increase the value (e.g. 1000).

    If this is not the point, we think it may need to collect the client logs for future analyzing when the issue reproduces.


    Best Regards,
    Alice Wang


    Please remember to mark the replies as answers if they helped. If you have feedback for TechNet Subscriber Support, contact tnsf@microsoft.com.


    Click here to learn more. Visit the dedicated forum to share, explore and talk to experts about Microsoft Teams.

    Tuesday, April 3, 2018 5:55 AM
  • Hey all,

    I would like to inform you that it seems like that I have just solved my problem with the ASMCU capacity.
    First of all, the KHI indicators are great.
    The only thing I mentioned there was the value: Packet Received Discarded:
    Our Servers are running in VMware and the following link helped me to reduce the amount of discarded packages.
    https://kb.vmware.com/s/article/2039495 --> Just for your information :)
    I have set up the values on one Server to max and it directly shows 0 discarded packages after a day. I will increase it gradually on the others too.

    Now back to the main problem.
    As I mentioned before this problem only occurs when s.o. from external establishes a connection to the ASMCU.
    I have just changed our Edge LB scenario to DNS LB. Before this change we had an unsupported HLB deployment. I have mentioned that the Media estbalishment with the ASMCU is now much faster than before and the Server did not show any kind of capacities failures after a lot of testing.

    So this seems to resolve my problem :) Thank you for your help!

    Greets and a nice weekend,

    Markus

    • Marked as answer by Markus_95 Saturday, April 7, 2018 2:33 PM
    Saturday, April 7, 2018 2:33 PM