PCoIP session freeze but rdp works in some special situations !

[omidgh13]

Hello All

i have installed a horizon view infrastructure with connection server and agent version 6.0.1 . everything was working well till i create some new pool for our developers . they are using Oracle tools and etc . we have more than 200 client in these pools , but about 10 percent of them get black screen randomly . i search too much and tried many things from timeouts , VGA , Firewall , ... but nothing works for me . then i started to focus on specific user and i got these :

1- if user start his desktop and work with normal apps , nothing happens . if he disconnect his session , he can get back and works again .

2- if user open Plsql ( or some similar apps )  and the apps crashes , if he disconnect the session and come back again , will get black screen . even if i End Task that crashed application , the same thing happens ! but if i logoff the session , no problem happens for next login !

3- in situation 2 , i can connect to that desktop with simple rdp ! and the system works fine . but i can not sign out that user with task manager !

so it seem that some situation make pcoip crashes . but even restating services or processes not worked for me . i have to restart that system ( it takes too long !) . one of my friend have the same problem with other versions . sometimes , when user disconnect the session , he cannot come back because of blank screen .

it is so painful for me . please help :smileyconfused:

( you can see these erro repeating after that crashe of pcoip :

07/13/2015, 16:44:14.353> LVL:2 RC:   0          COMMON :SOCKET_TRACE: tera_sock_socket_close() closed socket 8108 - currently 0 sockets.

07/13/2015, 16:48:00.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:01.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:02.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:03.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:04.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:05.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:06.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:07.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:08.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:09.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:10.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:11.295> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:12.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:13.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:14.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:15.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:16.296> LVL:0 RC:-504        MGMT_IMG :ERROR: tera_msg_queue_put failed on event 0x8 for PRI 0!

07/13/2015, 16:48:16.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:17.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

07/13/2015, 16:48:18.296> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

[dhargo]

Hello. I have exaclty the same issue and I've opened support request but VMware support engineers also can't find any solution or even sourse of the problem.

Do you have any new information how to make pcoip stable?

[roneng]

I have similar issues to what you describe with horizon view 6.1.

i had the same issues of some application crash the pcoip, and inly reboots fixes it.

i had multiple support cases, but with no resolution.

I even had a special dll created for me by vmware engineering to try and fix it , but that also did not work.

I also have issues where in the middle of a session to screen goes black and come back after ~20 seconds - also no help from support.

I must say that RDP does work, and also Citrix ICA, makes me wonder if PCoIP is not fully baked.

I would be happy to here if someone was able to solve these.

[dhargo]

Here is several optimizations which may affect the problem with pcoip:

1) Select "Auto-detect" for Displays and video memory in VM settings (Check: vCenter -> VM settings -> Video card -> Displays and video memory. There should be active "Auto-detect" settings. In my case I had active "Specify custom settings" and only 32 Mb video memory configured. So it can cause problems with black screen).

2) Configure VMXNET 3" as NIC for VM (Check: vCenter -> VM settings -> Network adapter -> Adapter type. It should be "VMXNET 3". In my case I had NIC "E1000". So it can cause problems).

3) Disable Screen Saver in Windows power options (open Windows -> Control Panel -> Power Options).

4) Disable IPv6 (Windows -> Control Panel - Network adapter)

I've just made these optimizations on Parent VM and next time I will reconfigure existing VMs one by one but it will take long time for me to reconfigure all VMs and get feedback if this settings can solve the plroblem.

If you can apply these settings and Recompose all your VMs then may be you can get result faster.

[TimothyBates]

VMware View 6.0.0 here, and I am seeing almost the same thing, but I do not see any apps crashing. User reconnects to an existing session, black screen with only the mouse cursor visible, cannot access the task manager or anything else.

It only recently started about a week ago. I am seeing this happening on multiple non-persistent pools, and none of them had been updated recently with a recompose before the issue appeared. I have a support case open and uploaded some logs, but I'm still awaiting a response.

-When it happens to a user, in the event log you do see an AGENT_PENDING_EXPIRED event on the VM in question about 30 minutes after the black screen connect. This is of course a system of the PCOIP connection was never really made

-We use a refresh after logoff environment with mostly non-persistent desktops. If you place a desktop in maintenance mode, it will still logoff after the four hour timer. However I cannot log the VM off through VMware View Administrator

-If I take a maintenance mode desktop, login to it via the vSphere console, and then try to reconnect, I receive "All available desktop sources for this desktop are currently busy"

-A reboot will clear the issue on the desktop

-We have a cluster running in EVC mode AMD Opteron™ Gen. 3 (no 3DNow!™). Most of the hosts in the cluster are AMD 6180 SE processors, but I notice the vast majority of the VMs with this issue are on hosts with newer 6344 processors.

-I am able to vMotion the VM to another host successfully, but it will not clear the issue.

Things that have not fixed the issue:

-Rebooting all of the View Connection servers

-Maintenance mode and rebooting ESXi hosts. Though this does seem to clear it up for a bit, but desktops will still start coming back with these black screens

-Pushing out a recompose with a new snapshot

[Masterkic2]

Same or really similar problem here with Vmware View 5.3.

We notice two warning with VMware Tools in the Windows events with one of our desktop having the problem.

[ warning] [vmusr:vmtoolsd] Error sending new-style capabilities: (null)

[ warning] [vmusr:hgfsd] Setting HGFS server capability failed!

But same error in the pcoip_server log

04/04/2016, 08:39:45.594> LVL:0 RC:-504        MGMT_IMG :mgmt_img_timer_perf_timer_cback() tera_msg_queue_put() failed on event 0x9 for PRI 0!

If someone get an answer from VMware please post it.

Thanks.