merse
Hot Shot
Hot Shot

Failed to connect to host (Template Deployment)

Hi,

When I deploy a template to a VM I'm getting a random success but a majority 'Failed to connect to host' error. I've removed and re-added from inventory, converted to VM and back many times, restarted VC services on the hosts but nothing seems to make a difference.

Is this a known problem or have I got something configured incorrectly somewhere? Just wanted to see if I was alone on this - everything else is running okay.

I'm running VC 2.5 64192 and ESX 3.5.0 64607.

Cheers,

merse

0 Kudos
20 Replies
dwalker-isi
Contributor
Contributor

I am getting the same issue trying to clone or add a new VM from the VI client connected to our VC console server. The tasks sit there queued for 15 to 20 minutes then time out with that error "Failed to connect to host".

I have seen other posts recommending a restart of the VMware services on the VC console server and/or a reboot of the VC console server. I have tried this and still have the cloning and 'create new vm' task queuing for several minutes and timing out.

I saw another mention of removing the licensing and readding it to the ESX server. I have not tried this yet but it is worth a shot.

Let me know if any of this helps.

Dan

0 Kudos
merse
Hot Shot
Hot Shot

Thanks for the response Dan, sounds like slightly different issues to me but indeed the licenses file is worth a go. Mine times out after around 10 seconds. I only have two ESX servers so I'm going to isolate each server and test to see if it's only one of them responsible... Good luck with yours!

0 Kudos
boydd
Champion
Champion

Have you tried restarting the management agents on the ESX side of the house? They try the deployment.

DB

VMware Communities Moderator

DB
0 Kudos
dwalker-isi
Contributor
Contributor

I restarted services but now ESX server is not responding and is in disconnected state on VC server console.

The server is running but I cannot reconnect the ESX server to the console.

Other ESX servers are still connected okay.

0 Kudos
boydd
Champion
Champion

Don't panic - it'll reconnect. One thing that I have learned with this product (and many other) is to be patient Smiley Happy

DB

VMware Communities Moderator

DB
0 Kudos
dwalker-isi
Contributor
Contributor

The vpxa service won't stay started but the mgmt-vmware service restarts ok. Still cannot get this ESX reconnected to console.

# service vmware-vpxa start

Starting vmware-vpxa:

# service vmware-vpxa status

vmware-vpxa is stopped

0 Kudos
admin
Immortal
Immortal

Can you look in the vpxa logs (/var/log/vmware/vpx/vpxa*.log) to see why it keeps crashing?

0 Kudos
dwalker-isi
Contributor
Contributor

Vmacore::InitSSL: doVersionCheck = false, handshakeTimeoutUs = 120000000

Another process is already running for this config file : pid=8263

Stub: Preference_GetBool called

Failed to initialize authd module functionality. Is vpxa already running?

Failed to do early initialization

0 Kudos
admin
Immortal
Immortal

That's probably not the root cause. Can you check out the other vpxa logs for a different errors at the end of the log?

0 Kudos
dwalker-isi
Contributor
Contributor

There were 10 log files vpxa-0.g to vpxa-9.log and they are all essentially the same except for time and so other numbers. I included log 0 and log 9 .

# more vpxa-0.log

Log for VMware VirtualCenter, pid=27939, version=2.0.2, build=build-50618, option=Release, section=2

Current working directory: /var/log/vmware/vpx

Initializing SSL context

Using system libcrypto, version 90701F

Vmacore::InitSSL: doVersionCheck

= false, handshakeTimeoutUs = 120000000

Another process is already

running for this config file : pid=8263

Stub: Preference_GetBool called

Failed to initialize authd module

functionality. Is vpxa already running?

Failed to do early initialization

# more vpxa-9.log

Log for VMware VirtualCenter, pid=28432, version=2.0.2, build=build-50618, option=Release, section=2

Current working directory: /var/log/vmware/vpx

Initializing SSL context

Using system libcrypto, version 90701F

Vmacore::InitSSL: doVersionCheck

= false, handshakeTimeoutUs = 120000000

Another process is already

running for this config file : pid=8263

Stub: Preference_GetBool called

Failed to initialize authd module

functionality. Is vpxa already running?

Failed to do early initialization

0 Kudos
admin
Immortal
Immortal

Just to confirm - you're using VC 2.0.2, right? Can you shut down the vpxa process (service vmware-vpxa stop), check that no vpxa processes are running, and then restart it (service vmware-vpxa start). Do you see the same problem?

0 Kudos
dwalker-isi
Contributor
Contributor

Yes 2.0.2 is the version. I did attempt to stop the service but it failed to stop it.

# service vmware-vpxa stop

Stopping vmware-vpxa:

It takes about 10 seconds to respond with that failed message.

0 Kudos
admin
Immortal
Immortal

What's the output for "ps awx | grep vpxa"? Can you try sending those vpxa processes a kill signal (kill -9 <processid>)? Might have to try a reboot if that doesn't work.

0 Kudos
dwalker-isi
Contributor
Contributor

A restart doesn't stop the vpxa service either but then it shows that it starts okay. Maybe time to reboot the ESX server? Or do you think something else is wrong that a reboot won't fix?

# service vmware-vpxa restart

Stopping vmware-vpxa:

Starting vmware-vpxa:

0 Kudos
dwalker-isi
Contributor
Contributor

# ps 8263

PID TTY STAT TIME COMMAND

8263 ? Z 165:30 [vpxa ]

# kill 8263

# ps 8263

PID TTY STAT TIME COMMAND

8263 ? Z 165:30 [vpxa ]

# kill -9 8263

# ps 8263

PID TTY STAT TIME COMMAND

8263 ? Z 165:30 [vpxa ]

0 Kudos
admin
Immortal
Immortal

Is that the only vpxa process? If it is a zombie it is strange that it is holding onto resources preventing other vpxa instances from running. Can you list the output from running "ps -elf | grep vpxa"

0 Kudos
dwalker-isi
Contributor
Contributor

# ps -elf | grep vpxa

4 Z root 8263 1 0 75 0 - 0 do_exi 2007 ? 02:45:30 [vpxa ]

0 S root 323 25748 0 75 0 - 926 pipe_w 08:49 pts/1 00:00:00 grep vpxa

0 Kudos
admin
Immortal
Immortal

Can you also post the output from running "cat /proc/8263/status"?

0 Kudos
admin
Immortal
Immortal

2 more commands for you to get output for:

ps -eopid,tt,user,fname,tmout,f,wchan | grep vpxa

cat /proc/8263/stat

0 Kudos