locked
Distributed Cache RRS feed

  • Question

  • I have a 2019 on prem 4 server farm. Two app servers and two WFE's all with 24gigs of dedicated ram in a VM environment. I have distributed cache running on only one server (second app server) with a cache size of only 1,229 mb's  . Which is below the 10% the remaining 22gigs of memory MS suggests. 

    My company uses a separate application for document management so for the most part the farm size is not very huge. My question though is two fold. A. Should DC even be running on the app server as the search service runs on them (the one caveat to that is that both app servers were set up as custom so they too do double duty as WFE's)? B. should I increase the size on the one or add a second box to the cluster (and maybe also increase the size)? Thanks everyone in advance for your advice. 

    Tuesday, July 21, 2020 10:48 PM

All replies

  • DC should be running on the FEs or dedicated servers, not any backend servers. We can't tell you if your FEs are properly sized nor if your DC size is sufficient -- that information is generally identified as usage of the farm begins.

    You should follow or use MinRole unless you have a specific reason not to.

    https://docs.microsoft.com/en-us/sharepoint/install/overview-of-minrole-server-roles-in-sharepoint-server


    Trevor Seward

    Office Apps and Services MVP



    Author, Deploying SharePoint 2019

    Author, Deploying SharePoint 2016

    This post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Tuesday, July 21, 2020 10:59 PM
  • Hi Ctoper,

    Please refer to Trevor's reply to have a check, I think it could be helpful.

    If you have any updates, please feel free to let us know.

    This “SharePoint Server 2019” Forum will be migrating to a new home on Microsoft Q&A, please refer to this sticky post for more details.

    Best regards

    Itch Sun


    "SharePoint" forums will be migrating to a new home on Microsoft Q&A !
    We invite you to post new questions in the "SharePoint" forums' new home on Microsoft Q&A !

    Wednesday, July 22, 2020 2:23 AM
  • Hi Trevor the mini role configuration is definitely a great idea. Unfortunately at the moment my company is migrating from 2013 and the end of this month is the cutover date. So I don't know if I can be doing too much time completely changing our topology. 

    But I do get your point they should not be running on the backend servers. As for the FE's being properly sized the company that did the analysis (before my time here) went over recommend MS specs for each box. Since we have no users in your opinion do you think it would be too much of an issue to add the two FE's to the cluster and remove the app server? Basically the ask is to streamline the site as page loads are slower. Also looking into IIS compression to improve page load times. 

    Thanks again for your help

    Wednesday, July 22, 2020 2:10 PM
  • Don't use IIS features not enabled out of the box. Certain ones aren't supported and that won't really help regardless as the content is dynamic and sourced from the SQL databases.

    MinRole is a simple conversion within Central Admin. You could make the FEs into a Custom role (which is what you have now) and run all services from there, but that will reduce performance.

    Remember that this isn't a cluster :-) There is no clustering services involved, they're referred to as farms.


    Trevor Seward

    Office Apps and Services MVP



    Author, Deploying SharePoint 2019

    Author, Deploying SharePoint 2016

    This post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

    Wednesday, July 22, 2020 9:42 PM
  • I used the Brotli compression as outlined in this MS article. https://docs.microsoft.com/en-us/iis/extensions/iis-compression/iis-compression-overview while my performance didn't go up in my Lighthouse test we did notice an improvement in page loads.

    You are correct about the farms. When I said cluster I meant the cache cluster.  Sorry sometimes tough to get across in just a message : ) In any case I did add the service to the two FE's and removed it from my app server. It went fine but this morning I woke up to numerous 6398 errors and saw the services went down on both boxes. This was also the case in my Stage environment where I did the same thing. It was no problem to restart them again but some articles I've read online have people stating this happens repeatedly to them. I will be monitoring them throughout the day but if it happens again tomorrow any suggestions on how to remediate the issue?

    Editor's Note: I followed your advise I changed the roles for my FE's to both also function as Distributed Cache and that has fixed the issue. Thanks for all of your help

    • Edited by ctoper11 Thursday, July 23, 2020 6:13 PM
    Thursday, July 23, 2020 2:19 PM