PCOIP disconnect/reconnect issues - VMware View 6.1

[ReneKarpati]

Our vdi users are having complains that, when working on their Windows 7 x64 vdi (dedicated), they sometimes get spontaniously disconnected and can't reconnect to their vdi. Only if we reboot/reset the vdi or if the users waits for some minutes.

I also had the same problems last saturday (10-01) on many occassions.

My Windows 7 vdi got disconnected and when i try to reconnect i get only a black screen and the message that the remote connection to the computer has timed out.

Sometimes i had to wait around 10 minutes before i could reconnect again.

In the pcoip logfiles on my vdi i see such messages:

SERVER :map_agent_to_tera: DISCONNECT_FOR_RECONNECT -> TERA_DISCONNECT_CAUSE_HOST_BROKER_RECONNECT

10/01/2016, 08:06:30.547> LVL:0 RC:   0       MGMT_SESS :Tearing down the session

10/01/2016, 08:06:30.547> LVL:2 RC:   0          COMMON :tera_pcoip_teardown_and_exit: Waiting for the session to teardown...

10/01/2016, 08:06:30.547> LVL:2 RC:   0 MGMT_PCOIP_DATA :OPEN: Got EVENT_RESET. Transition into BYE, pcoip data handle: 0

10/01/2016, 08:06:30.547> LVL:1 RC:   0        MGMT_DDC :mgmt_ddc_fcc_callback: queuing TERA_MGMT_FCC_EVENT_RESET [pri=0]

10/01/2016, 08:06:30.547> LVL:2 RC:   0      MGMT_VCHAN :app_reset: Session has been dropped. Forcing all channels to close!

10/01/2016, 08:06:30.548> LVL:2 RC:   0        MGMT_IMG :CODEC: State change from CODEC_CFG_EXCHANGE to CODEC_DISABLED

10/01/2016, 08:06:30.548> LVL:0 RC:   0        MGMT_IMG :mgmt_img_codec_reset: Detected System teardown.

10/01/2016, 08:06:30.549> LVL:2 RC:   0        MGMT_IMG :CODEC: State change from CODEC_DISABLED to CODEC_DISABLED

10/01/2016, 08:06:30.550> LVL:0 RC:   0        MGMT_IMG :mgmt_img_codec_reset: Detected System teardown.

10/01/2016, 08:06:33.972> LVL:1 RC:-504 MGMT_PCOIP_DATA :BYE packet not acknowledged, aborting session

10/01/2016, 08:06:33.972> LVL:2 RC:   0 MGMT_PCOIP_DATA :mgmt_pcoip_data_set_media_activation: URBoIP is not used

10/01/2016, 08:06:33.972> LVL:2 RC:   0          COMMON :SOCKET_TRACE: tera_sock_socket_close() closed socket 1700 - currently 0 sockets.

10/01/2016, 08:06:33.972> LVL:2 RC:   0       MGMT_SSIG :Request to reset session (PRI: 0)

10/01/2016, 08:06:33.972> LVL:1 RC:   0           VGMAC :Stat frms: R=000000/000000/096949  T=001984/106904/024216 (A/I/O) Loss=0.00%/0.00% (R/T)

10/01/2016, 08:06:33.972> LVL:2 RC:   0          COMMON :TERA_PCOIP: SESSION_EVENT=TERA_MGMT_SYS_SESS_EVENT_RESET, disconnect cause (0x105)

10/01/2016, 08:06:33.972> LVL:2 RC:   0          SERVER :server main: cb_notify_session_status called (mask 0x10) with tera_disconnect_cause (0x105)

10/01/2016, 08:06:33.972> LVL:2 RC:   0          SERVER :server main: TERA_MGMT_SYS_SESS_EVENT_RESET with disconnect cause (0x105).

10/01/2016, 08:06:33.972> LVL:1 RC:   0          SERVER :map_tera_to_agent_close_code: TERA_DISCONNECT_CAUSE_HOST_BROKER_RECONNECT -> PCOIP_AGENT_CLOSE_CODE_BROKER_INITIATED

10/01/2016, 08:06:33.972> LVL:2 RC:   0          SERVER :connection_closed: code = 2.

10/01/2016, 08:06:33.972> LVL:1 RC:   0           VGMAC :Session stats: Average TX=0.119944 average RX=0.0290176 (Mbps) Loss=0.01%/0.00% (R/T)

10/01/2016, 08:06:33.972> LVL:2 RC:   0        MGMT_SYS :Session closed remotely!

10/01/2016, 08:06:33.972> LVL:2 RC:   0        MGMT_SYS :********************************************************

0/01/2016, 05:47:20.915> LVL:2 RC:   0 MGMT_PCOIP_DATA :Memory stat: available physical memory = 1282MB/4095MB

10/01/2016, 05:47:22.248> LVL:2 RC:   0 MGMT_PCOIP_DATA :Tx thread info: round trip time (ms) =  26, variance =   2, rto = 128, last =  29, max =   0

10/01/2016, 05:47:41.768> LVL:1 RC:-504 MGMT_PCOIP_DATA :Unable to communicate with peer on PCoIP media channels (data manager ping timer expired)

10/01/2016, 05:47:41.768> LVL:2 RC:   0 MGMT_PCOIP_DATA :mgmt_pcoip_data_set_media_activation: URBoIP is not used

10/01/2016, 05:47:41.768> LVL:2 RC:   0       MGMT_SSIG :Request to reset session (PRI: 0)

10/01/2016, 05:47:41.768> LVL:2 RC:   0          COMMON :SOCKET_TRACE: tera_sock_socket_close() closed socket 1692 - currently 1 sockets.

10/01/2016, 05:47:41.768> LVL:2 RC:   0      MGMT_VCHAN :app_reset: Session has been dropped. Forcing all channels to close!

10/01/2016, 05:47:41.768> LVL:1 RC:   0        MGMT_DDC :mgmt_ddc_fcc_callback: queuing TERA_MGMT_FCC_EVENT_RESET [pri=0]

10/01/2016, 05:47:41.772> LVL:2 RC:   0        MGMT_IMG :CODEC: State change from CODEC_CFG_EXCHANGE to CODEC_DISABLED

10/01/2016, 05:47:41.772> LVL:0 RC:   0        MGMT_IMG :mgmt_img_codec_reset: Detected System teardown.

10/01/2016, 05:47:41.772> LVL:2 RC:   0        MGMT_IMG :CODEC: State change from CODEC_DISABLED to CODEC_DISABLED

10/01/2016, 05:47:41.772> LVL:0 RC:   0        MGMT_IMG :mgmt_img_codec_reset: Detected System teardown.

10/01/2016, 05:47:41.772> LVL:1 RC:   0           VGMAC :Stat frms: R=000000/000000/005109  T=000218/003305/004283 (A/I/O) Loss=0.00%/0.00% (R/T)

10/01/2016, 05:47:41.772> LVL:2 RC:   0          COMMON :TERA_PCOIP: SESSION_EVENT=TERA_MGMT_SYS_SESS_EVENT_LOST, disconnect cause (0x0)

10/01/2016, 05:47:41.772> LVL:2 RC:   0          SERVER :server main: cb_notify_session_status called (mask 0x80) with tera_disconnect_cause (0x0)

10/01/2016, 05:47:41.772> LVL:2 RC:   0          SERVER :server main: TERA_MGMT_SYS_SESS_EVENT_LOST.

10/01/2016, 05:47:41.773> LVL:2 RC:   0          SERVER :connection_closed: code = 3.

10/01/2016, 05:47:41.773> LVL:1 RC:   0           VGMAC :Session stats: Average TX=0.144996 average RX=0.0304322 (Mbps) Loss=0.00%/0.00% (R/T)

10/01/2016, 05:47:41.773> LVL:2 RC:   0        MGMT_SYS :Session lost!

10/01/2016, 05:47:41.773> LVL:2 RC:   0        MGMT_SYS :********************************************************

10/01/2016, 05:47:41.773> LVL:2 RC:   0        MGMT_SYS :********************************************************

10/01/2016, 05:47:41.773> LVL:2 RC:   0        MGMT_SYS :********************************************************

10/01/2016, 05:47:41.773> LVL:2 RC:   0        MGMT_SYS :********************************************************

10/01/2016, 05:47:41.805> LVL:2 RC:   0          SERVER :server mailbox: Stopping mailbox message loop

10/01/2016, 05:47:41.805> LVL:2 RC:   0          SERVER :server main: got terminate application message

10/01/2016, 05:47:41.805> LVL:2 RC:   0          SERVER :server main: exiting

10/01/2016, 05:47:41.805> LVL:2 RC:   0          SERVER :server cleanup: tearing down pcoip with agent disconnect reason code(0) and waiting

10/01/2016, 05:47:41.805> LVL:1 RC:   0          SERVER :map_agent_to_tera: DISCONNECT_GENERIC -> TERA_DISCONNECT_CAUSE_HOST_BROKER_GENERIC

10/01/2016, 05:47:41.805> LVL:0 RC:   0       MGMT_SESS :Tearing down the session

Can someone tell me how to troubleshoot these strange vdi connection issues?

I can add some of the full logfiles if that is useful , but the whole zip log bundle is around 700MB

In the View Admin console, global settings we use this:

Forcibly disconnect users: 720 minutes

Discard SSO credentials: After 15 minutes

[VirtualSven]

Why would you Forcibly disconnect users after 720 minutes if they have dedicated virtual desktops? It's not like resources will become available for other users. I would set that to "never" as it might be the cause of the "sudden" disconnects (after 12 hours maybe??).

Otherwise, this is a good start: https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=10306...

Or submit a support call with VMware and send the logs to them.

[MatthewDay]

Just some quick questions:

1. Does the disconnect event happen to everybody at the same time or just select users?

2. Is the problem occurring on multiple desktop pools or just one?

3. Do you have their desktops set to refresh on logoff? and if so, do you have anything that might be initiating a logoff command?

4. I see you are forcibly disconnecting users at 12 hours, is there any change their sessions are reaching the 12 hour mark?

5. Did the problem just start occurring? and if so, any major changes to the VDI infrastructure?

The first section of the logs you posted shows an event for: TERA_DISCONNECT_CAUSE_HOST_BROKER_RECONNECT, which could translate to the session was logged on from somewhere else and disconnected this session.

But, In the second section of logs, there is an event for: TERA_MGMT_SYS_SESS_EVENT_LOST, which usually means network connection loss.

Here is the KB with disconnect codes: PCoIP Disconnection Codes (2012101) | VMware KB

[ReneKarpati]

Hi Matthew

1. Three of our users made these complaints, maybe more experience the same problems but didn't notify us.

2. Problems occur on 3 different VDI pools (all dedicated vdi pools)

3. Automatically logoff after disconnect & Refresh OS disk after logoff are set to "Never"

4. We've changed the disconnect setting from 720 mins to Never, but problems unfortunately still occur

5. The users state these problems are occuring for many months now but finally notified us. No real big changes in the VDI infrastructure after the installation/configuration

Thanks for your input and help so far.

I will also submit a support call with VMWare + logfiles

[ReneKarpati]

Hi VirtualSven,

Thanks for your input.

We've used the 720 mins settings because otherwise the users will never logoff their vdi.

Changed this setting last week to "never" but unfortunately the problem still occurs.

We've checked the power saving settings on local system and vdi

(already has the setting never put the computer to sleep or turn off the display)

We will check the other points on the checklist to see if we missed something.

I will also post the support call with logs later this week.