This message was deleted.
# citrix-vad
s
This message was deleted.
b
Could be administratively hidden so the user could log back in elsewhere if it was a ghosted session
d
Is the disconnected session visible on the VDA?
j
Yes it's visible from the VDA, just doesn't show in Studio
@Benjamin Crill are you talking about that daft PowerShell method of hiding it?
b
Yeah, I know in larger environments the session can ghost and user can't log in at all. So admins will hide it until they can reboot the VDA so as not to disrupt other users
j
Sounds a bit advanced for this mob...the fact that it is holding the FSLogix profile locked would mean if they are doing that, then they're not helping
b
yeah that could be problematic
s
your sure the VDA registration is good? If you bounce the ctirix desktop service, does it reinitialize and then register and then show the session?
j
Some enterprising helpdesker logged the session out from the VDA as I was testing that. It was showing as registered though
s
dang. I've seen it the other way too where studio shows a bunch of sessions and being registered, but the VDA doesn't. Registration failed somewhere uncleanly.
j
It might make sense as this environment is also reporting issues with VDAs becoming unresponsive and unregistered....
s
oh boy, sounds like a fun one
j
I also have some users with 20min+ logon times and Director shows nothing that would account for that discrepancy....it's fun 24x7
d
Which VDA are they using? We saw both of these issues on VDA 2203 CU1. Sessions not showing up in Monitor (Director) and VDAs going unresponsive. After reviewing a few different discussion threads about this, we ended up adding these two registry keys to fix it. HKEY_LOCAL_MACHINE\SOFTWARE\Citrix\Ica\GroupPolicy "EnforceUserPolicyEvaluationSuccess"=dword:00000000 HKEY_LOCAL_MACHINE\SOFTWARE\Citrix\Reconnect "FastReconnect"=dword:00000000 We have since upgraded to VDA 2203 CU2 and left those reg keys in place.
More information....Upgrading Workspace to 2203 CU2 made this problem much worse. We went from 1-3 VDAs locking up per day to 5-10 per day. We found that users would start one session for the day and then for some reason they would start a second session within 10-15 mins. We're not sure if that was because they were roaming from one machine to another, which is typical at a credit union branch, or because their session would end abnormally. Anyway, the first session would not show up in Director and that's why they would start a second session. Of course their FSL profile was locked in the first session. Here's the kicker though....when their first session, which was in an Active state still on the VDA, would reach the Idle -> Disconnect timeout (4hrs) the session would attempt to disconnect and it would lock up the whole server! Existing sessions on the same server would lock up and no new sessions could log on. We could not gracefully reboot the server and had to hard power it off.
BTW - This is on Server 2019. This took us months to figure out. We tried upgrading the VDA, FSLogix, WEM, etc. along with the monthly MS hotfixes and nothing helped until we added those reg keys. Honestly, I'm not sure if both are required or just one was the fix, but our environment has been stable since.
b
@decamp both are required. Struggling since last summer with this issue...
j
This is Server 2019 too.
Sounds really ominous.....
These VDAs are on 1912 LTSR CU5
b
@James Rankin issue started with 1912 CU4 / 2203.0. Not fixed in 1912 CU4,CU5,CU6 and 2203 CU0, CU1, CU2. Workaround are the reg keys mentioned above.
j
Will give that a bash then. Is it fixed in 2206? Not that it matters for this customer as they are LTSR....
b
2206 has been affected as well if I recall correctly.
j
FFS
b
Check CVADHELP-20129. Took me 3 months, worst CTX Support experience ever.
g
Hmm... this sounds a lot like the issues I saw when Fslogix was acting up. The sessions were listed as disconnected on the servers, and active in Studio. Sometimes they might even disappear in Studio. Some servers eventually stopped responding alltogehter. This was with the latest version of Fslogix, you probably all know that story. I downgraded to Fslogix 2201 HF2 and corrected some missing exclusions in Defender ATP on the VDAs and the fileservers where the profiles are stored, and all problems have dissappeared. I've seen one single user having a locked profile in the week after I downgraded, as opposed dozens each day. Edit: We're running VDA 2203 here, can't remember exactly what build atm.
j
Hmmm....this customer is on an old (1909) FSLogix. Definitely seems as if the problem exists somewhere between FSLogix and the VDA though
g
I would guess it is worth testing 2201 HF2, if you have a QA environment at least.
j
Haha QA environment, I should be so lucky
😂 3
g
Same here really... I downgraded fslogix with a hail mary and a feable hope of success in production. 😛