vCenter connection lost

Mathias_Robichon · ‎2013-04-09

Hi folk,

During a POC Balance, we encounter a problem. The infrastructure is made up of 2 vCenter servers with the same configuration (from memory, one is vSphere 4.1u1 and the second 4.1u2).

The first one is working well. The second loses regulary the connexion during the gathering process. You are able to see the end og the logs file here :

2013-04-05 08:57:45,671 FINE Working on Virtual Machine name=0199412368-AMETIFFAX01 with mor=vm-95908

2013-04-05 08:57:45,671 FINE guestFamily for 0199412368-AMETIFFAX01=windowsGuest

2013-04-05 08:57:45,671 FINE guestGuestFullName for 0199412368-AMETIFFAX01=Microsoft Windows Server 2008 R2 (64 bits)

2013-04-05 08:57:45,672 FINE configGuestFullName for 0199412368-AMETIFFAX01=Microsoft Windows Server 2008 R2 (64 bits)

2013-04-05 08:57:45,672 FINE Using guestFamily to determine OS type for Virtual Machine: ${virtualMachine.getName()}

2013-04-05 08:57:45,672 FINE os type for 0199412368-AMETIFFAX01=WINDOWS

2013-04-05 08:58:25,211 SEVERE Cant resolve {0} to valid non-localhost IP address

2013-04-05 08:58:46,211 SEVERE Cant resolve {0} to valid non-localhost IP address

2013-04-05 08:58:46,211 SEVERE No valid hostname was found for this VM : 0199412368-AMETIFFAX01

2013-04-05 08:58:46,211 FINE Working on Virtual Machine name=1079045769-GRDFHYPBDDDEV with mor=vm-97129

2013-04-05 08:58:46,211 FINE guestFamily for 1079045769-GRDFHYPBDDDEV=windowsGuest

2013-04-05 08:58:46,211 FINE guestGuestFullName for 1079045769-GRDFHYPBDDDEV=Microsoft Windows Server 2008 R2 (64 bits)

2013-04-05 08:58:46,211 FINE configGuestFullName for 1079045769-GRDFHYPBDDDEV=Microsoft Windows Server 2008 R2 (64 bits)

2013-04-05 08:58:46,211 FINE Using guestFamily to determine OS type for Virtual Machine: ${virtualMachine.getName()}

2013-04-05 08:58:46,212 FINE os type for 1079045769-GRDFHYPBDDDEV=WINDOWS

2013-04-05 08:59:09,227 SEVERE Cant resolve {0} to valid non-localhost IP address

2013-04-05 08:59:30,227 SEVERE Cant resolve {0} to valid non-localhost IP address

2013-04-05 08:59:30,227 SEVERE No valid hostname was found for this VM : 1079045769-GRDFHYPBDDDEV

2013-04-05 08:59:30,474 SEVERE Error in Collection - COMMAND_EXECUTION_FAILED: Virtual Center Discovery Collection Interrupted for 172.30.101.100 Uncategorized exception occured during JMS processing; nested exception is javax.jms.JMSException: java.io.IOException - for id 103,232 type DISCOVERY element VIRTUAL_CENTER:172.30.101.100:443 Handle Name: 172.30.101.100

The full file is enclosed.

The errors SEVERE No valid hostname was found for this VM : 0199412368-AMETIFFAX01 are normal" because the VMs aren't is the same DNS domain. We got several. The last line is the problem.

Anybody has any hints ?

Regards,

Mathias

dmilliro · ‎2013-04-10

Hi Mathias,

In looking at the log file you sent and talking to engineering this is usually caused by the vCenter server closing our connection. Do you know if this vCenter server in question is under-provisioned? If there aren't proper resources on the end device our collections may fail.

Thanks, Daniel

Mathias_Robichon · ‎2013-04-12

Hi thanks a lot for your reply,

The vCenter server is loaded. Also, vmware reports a bug with the UI on port 443 when a lot of objects are configured. So, I guess that Balance uses port 443 with API and not the UI. Anybody could confirm that ?

If Balance uses only API, it could confirm a load problem with the vCenter.

Thanks a lot.

Mathias

dmilliro · ‎2013-04-12

Hi Mathias,

You are correct. Balance is talking to vCenter using API calls.

Let me know if you have any other questions.

Thanks, Daniel