Seen this issue cause some pain recent enough.
XenApp 6.5 HRP05
Published Desktop running W2008 r2 and Office 2013 using PVS 7.9 golden images
Received reports of users getting hung sessions. On reviewing system see it can’t be RDP’ed too, completely froze. Issue caused by write cache spilling over from RAM onto hard drive and filling hard disk.
What caused the write cache to blow up was chiefly offline address books being downloaded every time the user logged into outlook. As our environment is quite large these are in around the 500mb size range. With these systems being rebooted once a week and often times 10 logons on a system in a given day quickly noticed write cache growth across the board by Wednesday’s precipating reboots.
In order to fix we needed to disable download of the Offline address book when users’ accessed Outlook. We run in a cached mode in Exchange but was fine to disable the need to download.
Key is to add the registry hack in
and we were good to go
If so advised to exclude the following file from synchronizing:
As not required/serves no purpose in a UPM environment and contributes to profile bloat
Sometimes, with the best will in the world hardware will fail, and yes sometimes even “redundancy” eg RAID and what not won’t cover – if i had a quid for every instance i’ve seen failures for seemingly covered-across-the-board from a fault tolerant and something will fall down, whether its the controller or UPS causing a short circuit.
Either how just to give that extra bit of assistance to the friendly admin if something occurs to the user store for your profiles good to have checks in place from a software standpoint to limit potential corruption for a flaky file server.
- Edit the GPO / citrix policy that contains your Citrix profile management policy settings
- Go to Profile Manager | Advanced Settings
- Double-click the log off user if a problem is encountered policy
- Select Enabled
This will configure UPM to display an error message to users and then log users’ off. Although probably not appreciated at the time, in the long run they would (if they could remember to thank you:))
Very easy one this – one thing to note is if looking at edgesight session start detail for a user you will see a profile load time – this doesn’t just include profile copy down but also stuff like printer enumeration :
So to get the specific amount of time it takes to load a profile down using UPM:
I) RDP to the server you want to check the profile load time on
ii) In windows explorer navigate to c:\windows\system32\logfiles\userprofilemanager\ (unless you’ve specified a different folder for the UPM logs it will default to here
iii) click on “date modified” field and Double-click on the latest UPM log file
will default to loading the file in notepad – press CTRL+F and type ‘finished logon’
will arrow you down to a line similar to the following – note the time on the right-hand side is in seconds and that’s it – that’s how long it took UPM here to take down your profile
;6;10228;DispatchLogonLogoff: ———- Finished logon processing successfully in [s]: <0.53>.
Want to know how to find out how long it takes your citrix profile management profile has taken to be copied down on logon ? here’s how to check
First RDP to the server you logged onto with you administrative level privileges, if in an enterprise environment most likely will be a separate ID to your “every day” citrix user id
Navigate to the following folder (unless directly specified elsewhere via GPO will be the default location):
Launch the latest .log file
Do the search for the below search term
Will show in the above how long it took to copy down – as per example above was less than a second J
Numerous articles about this piece can be found but if shortly due to roll out as a published app or as part of a published desktop make sure if using citrix profile management to include the below list of folders for exclusion. Google chrome can cause major profile bloat