High CPU and Disk I/O when mapping Session Printers on W2K8R2 RDS

A client experienced Intermittent High CPU and Increased Disk I/O with SPOOLSV.EXE When Mapping TS User Session Printers on Windows Server 2008 R2 (Provisioned on Citrix Xenserver 6.0.2). A large amount of printers where mapped for each new user session. During the most logon extensive period in the morning the whole environment came to a grinding halt. After thorough investigation we noticed that with each client session starting heavy disk I/O was noticeable in the %system32%\spool\driver\3 folder. One could see the creation of a {GUID}-folder for each printer created during a session logon. The folder disappeared straightaway after the printer was mapped.

On Social Technet this issue was described in this thread.

As this behavior of mapping a printer in a user session and reinstalling each driver again is by design the resolution lay in the restriction of the number of mapped printers per user. Mapping one or two printer per user session resolved their performance issues.

2 comments

  1. fab says:

    Hello,

    I have exactly the same problem. Please could you explain me (if your remember) what is the resolution ? How and where have you mapped printers with GPO or using policies in citrix ?

    Thank you by advance.

    • Hi Fab,

      the resolution is in the restriction of the number of printers mapped per user at logon as stated in the last paragraph of this article.

      With this specific customer we used RES to provision the user environment. Since the problem is with the way Windows 2008 R2 handles printer driver mapping, it is not down to which way you map printers (GPO, Citrix GPO, RES, etc), but the number of printers you map per user.

      I just noticed, that the link in this article is dead. This is probably down to the fact that Windows 2008R2 is no longer supported.

      Kind regards,
      Misja

Leave a Reply to Misja Geuskens Cancel reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.