Storage Spaces direct and high CPU usage

John 76 Reputation points
2021-09-03T12:58:32.413+00:00

We have a 2 node Server 2019 cluster with the scale out file services role installed. We are using it to store FSLogix user profile disk. It's worked great up until recently, the CPU on the active node is very high causing the storage to run brutally slow. Trying to figure out what could cause this as our disk latency spikes to numbers in the seconds. We have an open ticket with MS, but no word or ideas as of yet.

Our S2D is stored on 4x intel 800GB Data Center SSDs and we have about 50-80 users signed into Remote desktop at a time. The only thing S2D is used for in RDS is the FSLogix User Profile Disk, which is essentially the Appdata folder.

Not sure if this is a storage issue and with it being software it's maxing out the CPU or if the CPU is maxed out causing the storage to suffer. Currently each cluster node has a E5-2670V3 CPU and we can add a 2nd if needed to each if the CPU is the bottle neck.

Any suggestions?

Thanks

Windows Server Storage
Windows Server Storage
Windows Server: A family of Microsoft server operating systems that support enterprise-level management, data storage, applications, and communications.Storage: The hardware and software system used to retain data for subsequent retrieval.
631 questions
0 comments

3 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Limitless Technology 39,381 Reputation points
    2021-09-03T18:19:40.837+00:00

    Hello @John

    In my experience it can be caused by High storage IOPS used by some applications or by users.

    Please also look at storage portal if can get some i/o matrix.

    to narrow down the issue please try to limit the users logins from 80 to 40.

    Please observe performance during off hours.

    If the reply was helpful, please don’t forget to upvote or accept as answer.

    1 person found this answer helpful.
    0 comments
  2. John 76 Reputation points
    2021-09-03T18:22:40.077+00:00

    It does seem to be fine when we limit the users to 40. Any ideas on why all of a sudden we were fine with 80 users then suddenly it maxes out at 40-45 before issues arrise?

    Thank you

    0 comments
  3. John Kay 21 Reputation points
    2021-09-24T15:24:17.493+00:00

    We seem to be getting an error in server manager now where the name of the Scale Out File Server and the Name of the cluster get Kerberos Security errors.

    Configuration refresh failed with the following error: The metadata failed to be retrieved from the server, due to the following error: WinRM cannot process the request. The following error with errorcode 0x80090322 occured while using Kerberos authentication: An unknown security error occured.

    This error seems to come and go with no rhyme or reason.

    Any suggestions?