locked
SQL requirements for WSUS 6 on WS2012? RRS feed

  • Question

  • Hello.

    We are implementing a new WSUS deployment on WS2K12 - approximately 15K clients.

    The topology is one upstream (root) WSUS server, with 2 downstream WSUS (replica) servers.

    All 3 WSUS servers are virtual machines.

    The replica servers will use WID for the database.

    The root WSUS server could either use WID or SQL 2008 R2 SP1 cluster, I've read that SQL Express isn't recommended. 

    Questions -

    1) Any reason to choose the SQL cluster over WID for the root WSUS server?

    2) Does WSUS require a dedicated instance on the SQL cluster, or can it use the default instance (with about 50 databases currently)?

    Thank you.

     

    Wednesday, July 9, 2014 3:40 PM

Answers

  • On Friday the root WSUS server ran out of disk space where the files are stored (D:\WSUS).

    It was initially configured as 40GB, now it's 80GB.


    I would say you have approved way too many updates.

    File space usage is determined by the number of approved updates.

    Only updates that can actually BE installed should be approved; superseded updates, for example, which will never be installed, have no value in being approved -- except to eat up disk space.

    Real world.... a WSUS server with an appropriate set of approved updates will use between 10GB and 20GB of disk space.


    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Tuesday, July 15, 2014 6:02 PM
  • We are implementing a new WSUS deployment on WS2K12 - approximately 15K clients.

    The topology is one upstream (root) WSUS server, with 2 downstream WSUS (replica) servers.

    Questions -

    1) Any reason to choose the SQL cluster over WID for the root WSUS server?

    Absolutely not.

    2) Does WSUS require a dedicated instance on the SQL cluster, or can it use the default instance (with about 50 databases currently)?

    It could use the default instance, but... it doesn't need to be on a remote SQL Server, and it surely doesn't need to be on a SQL Cluster, and I would argue that you'll get better performance using a dedicated instance of WID than you'll get sharing a cluster with fifty other databases of any type or size.

    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Friday, July 11, 2014 12:12 AM
  • At what number of clients would we want to change from WID to the SQL cluster?

    The **ONLY** reason to use a back-end SQL Cluster is because you're so functionally dependent on the services of WSUS that you need to have three 9s of availability on the WSUS environment.

    But without deploying Fault Tolerance on the front-end, it's absolutely pointless to worry about using a cluster on the back end.

    Furthemore, while NLB + SQLCluster is one methodology for achieving fault tolerance, I believe there's a much simpler solution:

    Upstream Server + Multiple Replica Servers.

    A very specific feature enhancement was made to WSUS v3.0 in 2007 to allow ANY replica server to be reconfigured to an Upstream Server with only a few mouse clicks. If you really need fault tolerance, deploy a replica server with no clients as a hot standby in another site or datacenter. It's much cheaper, and exponentially easier to setup, configure, and maintain than an NLB+SQLCluster.

    And let's be real.... how many hours/days a month do you actually use the WSUS server? Are you in a position where having the WSUS server offline for a day or so would be catastrophic to the BUSINESS. (We're not talking about an inconvenience in deploying security updates, we're talking about actual Loss of Revenue due to the server being offline.)

    For almost every organization, the answer is Zero Impact. And if the answer is Zero Impact, this scenario is NOT a candidate for fault tolerance when you can achieve the same thing with just one more VM as a clientless replica server.


    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Friday, July 11, 2014 12:18 AM
  • A follow-up question if you don't mind...

    On Friday the root WSUS server ran out of disk space where the files are stored (D:\WSUS).

    It was initially configured as 40GB, now it's 80GB.

    We haven't configured what I would consider a substantial number of products or classifications...

    Products

    • Office 2010
    • Office 2013
    • Windows 7
    • Windows 8
    • Windows 8.1
    • Windows Defender

    Classifications

    • Critical Updates
    • Definition Updates
    • Security Updates
    • Update Rollups

    Real world, how much free space should be configured for file storage?

    Thanks,

    Mike

     



    The classification that may cause you to use more disk space is the Update Rollups. Just make sure you do NOT approve ALL of those Update Rollups. Currently I am using around 89 gigs, but some of this is 3rd party patches using Patch Manager from solarwinds. So I would not be a "typical" real world example of disk space. I would recommend about 20-30% more than suggested in the manual. Just in case we get a large update (e.g. Windows 8.1 seemed to be the largest this year.) I hope this post helps a little bit.
    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Monday, July 14, 2014 5:29 PM

All replies

  • 1. Don't use Windows Internal Database for that big installatio, use you sql cluster instead.

    2. No it does't need an dedicated instance.

    Wednesday, July 9, 2014 6:50 PM
  • Thank you.

    At what number of clients would we want to change from WID to the SQL cluster?

    Thursday, July 10, 2014 3:19 PM
  • Actually you should not use wid in a production enviroment at all...

    A wid database can not be administrated as a sql database can and therefore you should not use it in a production inviroment.

    Thursday, July 10, 2014 10:28 PM
  • 1. Don't use Windows Internal Database for that big installatio, use you sql cluster instead.

    2. No it does't need an dedicated instance.

    TOTALLY disagree.

    On both counts.

    First... if the WSUS server NEEDS a remote SQL Server, then it NEEDS a dedicated instance.

    BUT... this scenario doesn't need a remote SQL Server, much less a CLUSTER!

    15K clients is about 15% of the capacity of a WSUS server... based on =2007= hardware availabilities!

    Except, the upstream server won't have 15k clients; it will have TWO -- the two downstream servers. (Okay, maybe some LOCAL clients as well; and in that case, there's merit in deploying a local replica server to handle local clients, in which case the USS will have *three* clients.)

    In any event, use the Windows Internal Database on THIS USS.


    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.



    Friday, July 11, 2014 12:08 AM
  • We are implementing a new WSUS deployment on WS2K12 - approximately 15K clients.

    The topology is one upstream (root) WSUS server, with 2 downstream WSUS (replica) servers.

    Questions -

    1) Any reason to choose the SQL cluster over WID for the root WSUS server?

    Absolutely not.

    2) Does WSUS require a dedicated instance on the SQL cluster, or can it use the default instance (with about 50 databases currently)?

    It could use the default instance, but... it doesn't need to be on a remote SQL Server, and it surely doesn't need to be on a SQL Cluster, and I would argue that you'll get better performance using a dedicated instance of WID than you'll get sharing a cluster with fifty other databases of any type or size.

    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Friday, July 11, 2014 12:12 AM
  • At what number of clients would we want to change from WID to the SQL cluster?

    The **ONLY** reason to use a back-end SQL Cluster is because you're so functionally dependent on the services of WSUS that you need to have three 9s of availability on the WSUS environment.

    But without deploying Fault Tolerance on the front-end, it's absolutely pointless to worry about using a cluster on the back end.

    Furthemore, while NLB + SQLCluster is one methodology for achieving fault tolerance, I believe there's a much simpler solution:

    Upstream Server + Multiple Replica Servers.

    A very specific feature enhancement was made to WSUS v3.0 in 2007 to allow ANY replica server to be reconfigured to an Upstream Server with only a few mouse clicks. If you really need fault tolerance, deploy a replica server with no clients as a hot standby in another site or datacenter. It's much cheaper, and exponentially easier to setup, configure, and maintain than an NLB+SQLCluster.

    And let's be real.... how many hours/days a month do you actually use the WSUS server? Are you in a position where having the WSUS server offline for a day or so would be catastrophic to the BUSINESS. (We're not talking about an inconvenience in deploying security updates, we're talking about actual Loss of Revenue due to the server being offline.)

    For almost every organization, the answer is Zero Impact. And if the answer is Zero Impact, this scenario is NOT a candidate for fault tolerance when you can achieve the same thing with just one more VM as a clientless replica server.


    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Friday, July 11, 2014 12:18 AM
  • Actually you should not use wid in a production enviroment at all...

    Where, Pray Tell, did you get this CRAP from?

    A WID database can not be administrated as a sql database can and therefore you should not use it in a production environment.

    A WID database can be easily administrated (sic) (the word is administered) using SQL Server Management Studio, and WSUS, SharePoint, and several other applications have been doing so IN PRODUCTION ENVIRONMENTS for many years.

    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    Friday, July 11, 2014 12:23 AM
  • Henrik & Lawrence,

    Thank you both for your responses.

    We have implemented the WID database on the root and replica servers.

    Mike

    Monday, July 14, 2014 1:05 PM
  • A follow-up question if you don't mind...

    On Friday the root WSUS server ran out of disk space where the files are stored (D:\WSUS).

    It was initially configured as 40GB, now it's 80GB.

    We haven't configured what I would consider a substantial number of products or classifications...

    Products

    • Office 2010
    • Office 2013
    • Windows 7
    • Windows 8
    • Windows 8.1
    • Windows Defender

    Classifications

    • Critical Updates
    • Definition Updates
    • Security Updates
    • Update Rollups

    Real world, how much free space should be configured for file storage?

    Thanks,

    Mike

     



    Monday, July 14, 2014 2:06 PM
  • A follow-up question if you don't mind...

    On Friday the root WSUS server ran out of disk space where the files are stored (D:\WSUS).

    It was initially configured as 40GB, now it's 80GB.

    We haven't configured what I would consider a substantial number of products or classifications...

    Products

    • Office 2010
    • Office 2013
    • Windows 7
    • Windows 8
    • Windows 8.1
    • Windows Defender

    Classifications

    • Critical Updates
    • Definition Updates
    • Security Updates
    • Update Rollups

    Real world, how much free space should be configured for file storage?

    Thanks,

    Mike

     



    The classification that may cause you to use more disk space is the Update Rollups. Just make sure you do NOT approve ALL of those Update Rollups. Currently I am using around 89 gigs, but some of this is 3rd party patches using Patch Manager from solarwinds. So I would not be a "typical" real world example of disk space. I would recommend about 20-30% more than suggested in the manual. Just in case we get a large update (e.g. Windows 8.1 seemed to be the largest this year.) I hope this post helps a little bit.
    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Monday, July 14, 2014 5:29 PM
  • On Friday the root WSUS server ran out of disk space where the files are stored (D:\WSUS).

    It was initially configured as 40GB, now it's 80GB.


    I would say you have approved way too many updates.

    File space usage is determined by the number of approved updates.

    Only updates that can actually BE installed should be approved; superseded updates, for example, which will never be installed, have no value in being approved -- except to eat up disk space.

    Real world.... a WSUS server with an appropriate set of approved updates will use between 10GB and 20GB of disk space.


    Lawrence Garvin, M.S., MCSA, MCITP:EA, MCDBA
    SolarWinds Head Geek
    Microsoft MVP - Software Packaging, Deployment & Servicing (2005-2014)
    My MVP Profile: http://mvp.microsoft.com/en-us/mvp/Lawrence%20R%20Garvin-32101
    http://www.solarwinds.com/gotmicrosoft
    The views expressed on this post are mine and do not necessarily reflect the views of SolarWinds.

    • Marked as answer by Daniel JiSun Thursday, July 17, 2014 6:14 AM
    Tuesday, July 15, 2014 6:02 PM