This message was deleted.
# _general
s
This message was deleted.
n
If the content has already begun downloading, it's a losing battle, IME. I've tried stopping all services, and then deleting the contents of that directory, along with the catalog file, and the Feature Updates still shows up in Windows Update, and attempts to install at next runtime. This is outside of Citrix, and not using WSUS, so it may not be of any help.
a
thats what i was thinking too. i just setup a PDQ Inventory group to show any VM's showing higher than 1909 to keep an eye on things. I do have a script in there to run / apply windows updates so i could just do that off hours on not in use VM's so if they DO have it, it handles it off hours at least
n
Going to 2004, though... eww.
a
ya not a big fan of this either.. i actually didnt pay too close attention to 2004 since i never intended to go to it. any major things to watch out for? maybe the upgraded ones i just jump in and push them to 20H2 or 21H1
n
Just that it's not a Fall release, so support is limited. I just went from 1909 to 20H2, personally.
The updated Start menu is nice.
a
nice! i did push 20H2 to one physical machine and ran into all kinds of drive mapping issues so got scared off
but ya i knew at the very least these 50 or so (or we'll see how many) i need to update before they are end of life end of this year now
n
Yeah, we literally just finished our 1909 rollout on the physical side (thanks to another team dragging their feet), and now we have to plow right to 20H2. They don't like it, but it's their own damn fault.
On the Citrix VDA side, the only downside is that 20H2 consumes more resources.
a
yup exactly!
that sucks....how bad?
n
Still waiting for Go-EUC or somebody else to run another resource consumption test. It's been a while since the last one.
I have no hard numbers, but I bumped my VDAs from 5GB to 8GB because of how many I saw in testing maxing out.
a
we have 8gb ram as a min for all of ours already actually
n
How many CPU? Ours run 2 vCPU for our 'small' offering, which seems to be fine. Very basic workflow, though. Few hardcore/heavy hittersa.
a
1 vCPU / 4 core
developers are 1 vCPU / 8 core
n
We're 1/2 for small, and 1/4 for large.
This will likely change during our next hardware refresh.
a
cool thanks for the input! so other than resources no real issues on the citrix setup for 20h2 so far?
n
Oh, 20H2 also brings new methods for disabling Windows Defender, FWIW. There's another thread on this from last week if you're interested. Can't just disable it via GPO anymore and have it work.
That's the only real gotcha that I can think of.
a
cool thanks!
we actually just switched TO Defender lol
n
We did have an issue with about a dozen users who couldn't open their 365 mailbox (said 'need password' at the bottom), but I figured out a resolution for them. Never saw that in any testing whatsoever.
a
not my call
n
I'll be hitting you up at some point, then. We're switching from Trend to Defender next year.
a
interesting.....what did you find w/ that? we had that a year ago and found a reg workaround but wonder if you have something etter
sounds good. we were w/ Cylance before
n
Never found the source of the issue, but the resolution was as simple as deleting a registry key
I have a "Repair Outlook" shortcut in the Start menu for the users, so they can fix it themselves.
a
which key was it? we had to Add a EnableADAL dword (0) and DisableADALatopWAMOverride dword (1) to fix it for us
n
HKCU:\Software\Microsoft\Office\16.0\Common\Identity
I also kill HKCU:\Software\Microsoft\Office\16.0\Common\Licensing, because of a prior issue.
a
oh you delete the whole Identity key? and when that recreates itself you dont just get the issue back?
n
Those two keys busted our shit when I was testing.
Yep, allowing it to recreate Identity resolved it for us.
a
so probably some entry underneath Identity was bad for you obviously then
for us its w/ fresh systems even so maybe something different but the keys i mentioned were the trick for us so i just set those w/ WEM
n
Yeah. It was prompting for a password in Outlook, but you couldn't actually enter any creds. Real weird. So, right on par with O365.
a
and intune for our phyiscal laptops now
n
This only happened with existing profiles, and only after the 20H2 upgrade. New profiles have been fine.
The weirdest part for me is this:
a
ahhh weird. ya for us you just clicked it and it went through on its own but would come back, and back, and back...
super weird!
good to know when i try 20h2!
n
wait, I'm thikning of a different issue now
My Prod Citrix Desktop had that "Your OST cannot be opened" blah blah error.
I then logged into my QA desktop - which uses a totally separate FSLogix store... same issue, on the same day.
No chance that's a coincidence.
a
ahh i gotcha. we're still onprem and online mode but JUST starting moving mailboxes out the cloud so ost's will become a thing soon here
n
haha, oh you are in for a treat.
I miss Exchange on-prem online mode. Fuck caching.
a
oh i know! we've warned about going back to os's as much as we are able but lost that battle
i think they've minimized the ost nightmares of the past w/ time thinking ahh it wasnt THAT bad was it? lol
n
That's why I added that Repair Outlook shortcut. Somebody's OST is corrupt, they can fix it in seconds. We dont roam search, either.
a
so your repair shortcut is it JUST killing those 2 keys you mentioned ? assuming it recreates the ost then
n
Hold on, I'll share
a
awesome thanks
n
ResetWEM.ps1
a
nice thanks man i'll take a gander at them
n
One for Outlook, one for Teams, and one for WEM (figured I'd throw that in there, just for the hell of it, in case I needed a user to run a RunOnce task that they had previously run).
The Teams one, I also copy my base json file as part of that script.
a
nice thanks man! so the outlook one is for corrupt osts.....what are the teams/wem ones for ?
n
The Outlook one is for corrupt OST and that weird issue we saw with 20H2 that I just mentioned. Teams would be for anything, really. We haven't seen any issues at all, but if Teams doesn't load or something, rip it out and relaunch Teams, done. WEM, removes the tracking in the user's registry, so the next time they login, it looks like their first login to WEM, so any RunOnce task will run again.
a
very nice! those will be handy to have i'm sure! thanks for sharing!
now if MS could make the post feature upgrade reboot NOT take 45-60 friggin min we wouldnt be so terrified to apply them!
n
I can't take the credit. Saw this posted elsewhere on Slack a while back, so I modified it a bit to suit my environment.
a
not a problem! all a big community to share and tweak to our own setups !
n
I do think that 20H2 upgraded quicker than previous FUs, but I could be making that up in my head.
a
i think in my testing i found that post install reboot was in the 35-40 min range so maybe a tad faster but disappointing that takes so long
n
One of the benefits of being a non-persistent PVS shop, I guess.
a
yup for sure
so i was able to use a powershell scanner w/ PDQ to give me a dynamic list of any desktop that shows this entry (to give me an idea how bad this is gonna get...) and i have about 200 of them so not great. thinking since 2004 is EOL in December anyways i might as well just start going to 20H2......i wonder to save time if it would work to just do the 20H2 (which i have more control over w/ a deployment script while its in this pending/reboot to finish state for 2004 or if that would cause an issue.
n
IMO, I'd be leery of moving right to 20H2 with that status. I'd probably be inclined to let it roll with 2004 and plan accordingly for a move to 20H2 soon thereafter.
BTW, I love that you highlighted that text for me... lol
a
ya i'm running a test going direct to it just to see what happens on a test machine. taking a looooong time though so i think you're right. and that screenshot i actually had marked up for our help desk so just re-used haha
n
In that case, you need a few arrows pointing to the box...
a
and make everything outside of the box dim and flashing arrows