0

We recently increased our user amount on our RDS 2012 R2. We are starting to experience some issues.

We have 4 Windows 2012 R2 RDS Servers and 1 Domain Controller/File server also running Windows Server 2012 R2.

Every RDS server has +- 10 users, At Max 15.

Every now and then when a users logs on the GPO fails to map the network drives. After logging off and login again it might be working again, the user just has to be 'lucky'. The event log shows the following event:

Event ID: 4098
The user 'Y:' preference item in the 'RDS2012 {AAFF7C11-7514-4CC3-BBC6-C00BBE718106}' Group Policy Object did not apply because it failed with error code '0x80070008 Not enough storage is available to process this command.' This error was suppressed.

We try to connect to the shares using the FQDN.

In a very few cases i had users trying to logon onto the RDS and receiving an error:

The remote procedure call failed and did not execute.

Waiting a few minutes and the user is able to logon just fine.

A third issue is when users are working in an Microsoft Office application they suddenly get a messagebox: 

Disk is full or too many files are open

It looks like there is a perfomance issue. But after monitoring it, it seems fine.

RDS Servers:

  • CPU usage is around 20-40% most of the time
  • RAM usage is at 45-50% (7.7GB/16GB)

Domain Controller/File server: - CPU usage is around 20-50% - RAM usage is at 70-80% 3.2GB/4GB

I have also checked for the amount of open handles in taskmgr and its a total of +- 150k handles. I think this shouldnt be a problem either?

Does anyone have an idea what might be causing these issues?

Edit:

The drive mappings are partially fixed with running them in the user context (configurable in GPO). Now the network drives are added. Sometimes they still are still marked with a red X in 'This PC' but double clicking it fixes the problem in 99% of the cases.

The Disk is full or too many files are open warning message still pops up in Office applications. This is comming from the Auto-save feature.

I ran a Process Monitor and looked at the moment the errors occurred and noticed 2 (strange?) things.

1. While doing a WriteFile on the pipe services on the domain controller/file server the return value is always INSUFFICIENT RESOURCES.

PIPE Service return value

2. The network directory it's trying to read/write the auto-save file from/to has a double backslash in its name. Opening that exact path in windows explorer gives an error. After changing the double backslash to a single backslash i can open the directory just fine.

Double slash in network path

I checked the GPO redirections and auto-save location in the Office applications and none of these contain a double backslash.

eKKiM
  • 1,483
  • 9
  • 22
  • Did you see this? http://serverfault.com/questions/647247/cant-access-share-sometimes-and-some-strange-performance-issues – pat o. Apr 07 '16 at 12:43
  • 1
    Yes, used the FQDN from the start. – eKKiM Apr 07 '16 at 12:45
  • Are you using "User Profile Disks" and does that user have enough space within their DIsk? – Drifter104 Apr 07 '16 at 12:57
  • No, we do not use User Profile Disks – eKKiM Apr 07 '16 at 12:59
  • Did you tried with a user to flush is profile, to be sure it's not a old setting that got in is profile, and office was installed with a configuration file ? (or you deploy office setting by gpo, let us know) – yagmoth555 Jun 07 '16 at 16:21
  • Yes i have tried flushing the user profile. Office settings are deployed trough GPO. Today at the starting of the day, while alot of people are logging in, we had the "The remote procedure call failed and did not execute." issue with a few users again. – eKKiM Jun 07 '16 at 16:24
  • You might want to disable any AV/firewall/filtering type of software on both ends and test; make sure your server is up-to-date - it might be a bug in OS; – strongline Jun 10 '16 at 15:42
  • We already de-installed the anti virus on all hosts which could influence this (DC and RDS) there is no firewall except Windows firewall between the nodes. All hosts are up to date.. – eKKiM Jun 10 '16 at 16:11
  • What is your disk usage on your DC? How many volumes do you have? Are your drives configured as a RAID? Do you have open drive slots on your server? – user5870571 Jun 13 '16 at 03:44
  • If you ruled out network and disk quota issues, you might want to check out this article: https://community.spiceworks.com/how_to/128605-rpc-issue-when-connecting-to-unc-root-share – J-M Jun 13 '16 at 00:31
  • I think I've run across this issue before. Has this been resolved. RDS = Remote Desktop Service? So is this an RDP to the server? Are you using HA, replication and/or NLB? – JaredW82 Jul 25 '17 at 19:22
  • There is a linux-based load balancer for the connection to the RDP servers. Besides that there is no HA or Replication. – eKKiM Jul 26 '17 at 17:17

0 Answers0