Hi,
(I have logged this as a support case w/VMware, just mentioning here in case anyone can help)
*****
<Company> are running a Horizon View 6 Cloud Pod Architecture.
4x Connection Servers in the primary site/Pod
4x Connection Servers in the secondary site/Pod
Connection Servers are not linked to Security Servers but using F5 APM, so are running in Direct Mode.
Virtual Desktop guest OS: Windows 8.1, Horizon 6 agent v6.
Pools are configured as Dedicated, due to full-clone/persistent desktop requirements.
Issue:
VIPA is showing intermittant timeouts for inter-pod brokering tasks, which seems to result in a user account being entitled to a desktop on *BOTH* sites. Once this occurs, the user cannot log in as the pools are Dedicated and a double-up of entitlements is not allowed. This is remediated by unassigning the user from a desktop on one of the sites, or (if that returns an "internal error") deleting one of the desktops.
Example errors:
Timeout example – intermittent full timeout:
(SESSION:24a7_***_8aad) LMV: Secondary-CS-02: Target: SITE=875cfb18-3c10-43df-add1-a7020fd07646 (total time: 20002ms)
(SESSION:24a7_***_8aad) LMV: Secondary-CS-02: (Primary-CS-03.company.local/10.106.39.42:8472-17:Failed:connect=0ms:send=2ms:server=-ms:total=10004ms)
(SESSION:24a7_***_8aad) LMV: Secondary-CS-02: (Primary-CS-04.company.local/10.106.39.43:8472-14:Failed:connect=0ms:send=1ms:server=-1ms:total=10000ms)
(SESSION:24a7_***_8aad) LMV: getUserInfo request timed out. Overall request timeout: 20000. Subrequest timeout: 10000
Timeout example - timeout to one CS
<VipaServerSlowThreadPool2-20> [LinkedModeDesktopsHandler] LMV: Checking global entitlements
[VipaReply] (SESSION:54a5_***_6260) LMV: setUserSettings request timed out. Overall request timeout: 20000. Subrequest timeout: 10000
[SetUserSettingsVipaReply] (SESSION:54a5_***_6260) LMV: Unexpected number of setUserSettings reply messages
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Pushing global user settings from site: 875cfb18-3c10-43df-add1-a7020fd07646, pod: Primary_Pod
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Setting user settings... Scope: ANY, site: 875cfb18-3c10-43df-add1-a7020fd07646, pod: Primary_Pod
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Setting user settings... Scope: SITE, site: 875cfb18-3c10-43df-add1-a7020fd07646, pod: Primary_Pod
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Setting user settings... Scope: LOCAL, site: 875cfb18-3c10-43df-add1-a7020fd07646, pod: Primary_Pod
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Storing global settings
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Making remote setUserSettings request to 1 targets
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: Target: SITE=6a1e8b0e-aad1-4d9a-bc38-098f2fdc816f (total time: 20002ms)
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: (Secondary-CS-03.company.local/10.185.40.43:8472-8:Failed:connect=0ms:send=1ms:server=-1ms:total=10003ms)
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: Primary-CS-01: (Secondary-CS-04.company.local/10.185.40.42:8472-7:InProgress:connect=0ms:send=1ms:server=-1ms:total=-1ms)
[Breadcrumbs] (SESSION:54a5_***_6260) LMV: setUserSettings request timed out. Overall request timeout: 20000. Subrequest timeout: 10000
[LinkedModeDesktopsHandler] LMV: Desktop launched on pod Primary_Pod, site: 875cfb18-3c10-43df-add1-a7020fd07646, pool: Desktop_Pool(Desktop1)
Connection Servers all have 4vCPU/10GB memory. Utilisation of the Connection Servers is low/basically idle.
The Virtual Desktop guests do have an issue with CPU utilisation, due to an on-going issue with Internet Explorer. On a session disconnect, the virtual desktops lock out at 100% CPU utilisation.
VMware KB 2080760 details steps to increase the timeout values for VIPA requests. However, the two timeouts in particular (pae-RequestTimeout and pae-SubRequestTimeout) are noted as "VMware recommends that you do not change these attribute values".
Seeking advice as to whether we should increase the timeout values by 10 seconds each - pae-RequestTimeout from 20s to 30s, pae-SubRequestTimeout from 10s to 20s.
Inter-site latency is 10ms. No packet loss/networking issues we are aware of.
Any suggestions?
Thanks :-)