This message was deleted.
# _generals
Slackbot
08/26/2021, 7:52 PMThis message was deleted.
👀 1
n
Nathan Shaw
08/26/2021, 9:29 PMHave seen similar issues on physical Windows PCs. Not sure cause, offloading or something I guess. ESX I'd be making sure I had the VMXNET3 NICs, and the up-to-date VM tools.
i
Isaac L
08/26/2021, 10:09 PMHi Nathan, I updated vmware tools after the problem started about a month ago. It unfortunately did not change the behavior. Some days are better than others.
s
Steven Lemonier
08/27/2021, 7:20 AMMaybe AV?
n
Nathan Shaw
08/27/2021, 7:22 AMHow's big’s the file you're transferring?
w
Webster Webster
08/27/2021, 12:41 PMVMXNET3 NICs?
Receive Side Scaling Enabled?
All the Offloading Advanced properties left at the default of Enabled?
i
Isaac L
08/29/2021, 4:49 PMFigured out it is a problem with the mcs temporary cache. Removing it and just having the temporary disk seems to fix my proble. Still too soon to tell but will be testing this week.
Isaac L
08/30/2021, 6:29 PMSo far testing has shown that removing temporary cache has NOT fixed our issue. Will be working with Citrix Consulting later this week to see what we can find.
n
Nathan Shaw
08/30/2021, 9:14 PMHi @Isaac L how about the other things mentioned on here? Any insights?
i
Isaac L
08/30/2021, 9:21 PMHi @Nathan Shaw same behavior when using e1000e as when using vmxnet3. I was transferring a 4gb file and consistently freezing up the servers that have a temporary memory cache size of 2gb.
I then created a new catalog that does not use the temporary memory cache. I have not been able to reproduce the freezing with file transfers but it occured on one of the new servers during production today. I am going to put in some more AV exclusions on the VDA side and hope that helps.
Isaac L
08/30/2021, 10:04 PMReceive Side Scaling is disabled
Isaac L
08/30/2021, 10:05 PM@Webster Webster all the offloading advanced settings are default enabled. Do you recomend disabling?
n
Nathan Shaw
08/30/2021, 10:08 PM@Isaac L do you have the MCS disk cache also? How big? I'd also be looking at my storage system stats when this issue happens, to see if disk growth is slow (eg if thin provisioned). Not clear from your comments if this is a new issue, or only happening on certain builds etc.
w
Webster Webster
08/30/2021, 10:13 PMI do NOT recommend disabling
Webster Webster
08/30/2021, 10:18 PMRSS should be enabled. RSS gives TCP one thread per vcpu. In your case 6 threads. Vmware and Microsoft recommend enabling RSS.
i
Isaac L
08/30/2021, 10:23 PM@Nathan Shaw I do have a temporary disk of 125gb that is thin provisioned. I am wondering if thick may be a better option?
Isaac L
08/30/2021, 10:31 PM@Webster Webster i will test with RSS enabled thank you for the tip.
Isaac L
08/30/2021, 10:36 PM@Nathan Shaw this problem started for me 1 month ago. About a month before the freezing started we implemented FSLogix and profile directory mirroring/syncing. We use Citrix UPM and FSLogix for office.
w
Webster Webster
08/30/2021, 10:49 PMhave you read this?
https://kb.vmware.com/s/article/2008925
i
Isaac L
08/30/2021, 11:04 PM@Webster Webster oops, my power is set to balanced. Cpu rdy i have looked at and have my hosts at a 1:1 vcpu to pcpu ratio. Thanks for the article fingers crossed this will alleviate my issues or at least help a little.
w
Webster Webster
08/30/2021, 11:26 PMi have a gpo that sets all vms to high perf power plan
i
Isaac L
08/31/2021, 9:37 PM@Nathan Shaw @Webster Webster today I found some articles and discussions mentioning that the mcs io driver is fundamentally broken in 7.15 ltsr. Tomorrow we will be rolling out a catalog that does not use temporary disk. 🤞fingers crossed this will fix our issues.
w
Webster Webster
08/31/2021, 9:38 PMOh, I had forgotten about that issue
i
Isaac L
09/02/2021, 5:27 AMNot using the mcs io driver fixed our issue.
Going to upgrade to 1912 and test out the mcs io as I've read that it actually works and is an elegant solution for speed and non-persistence.
5 Views