FJ1200
Enthusiast
Enthusiast

Is a v5.0 Update Mismatch causing "Timed waiting for vpxa to start"?

Morning all

A customer has a vCenter 5.0 U1 managing ESXi 5.0 U2 hosts.  One host has had to be removed from the cluster and needs adding back in, but now we get the "Timed waiting for vpxa to start" error message when we try.  Personally I think the Update mismatch is causing the problem even though we're not going from esxi 5.1 to a vC 5.0, the differences are still there. Am I correct?  I haven't seen this error message before until now, and have checked all the possible reasons I've come across.   

Tags (4)
0 Kudos
5 Replies
FJ1200
Enthusiast
Enthusiast

Just to add

My vpxalog has entries like as seen in the post: Vcenter 5.1 Timed waiting for vpxa to start:

/var/log # cat vpxa.log | grep SOAP

--> while parsing SOAP body

--> while parsing SOAP envelope

2015-07-15T07:28:08.689Z [74468B90 error 'SoapAdapter'] Unsupported namespace "urn:vpxa3" in content of SOAP body

--> while parsing SOAP body

--> while parsing SOAP envelope

/var/log # cat vpxa.log | grep error

2015-07-15T06:58:05.751Z [74405B90 error 'Default'] [VpxVmomi] Exception while sending fault: Broken pipe

2015-07-15T07:28:08.652Z [74489B90 error 'SoapAdapter'] Unsupported version URI "urn:vpxa3/2.0"

2015-07-15T07:28:08.652Z [74489B90 error 'SoapAdapter'] Unsupported version URI "urn:vpxa3/2.0"

2015-07-15T07:28:08.689Z [74468B90 error 'SoapAdapter'] Unsupported namespace "urn:vpxa3" in content of SOAP body

I think it's the same issue even though it's just Update levels.  Any ideas anyone?

<Update>

I've been given the go-ahead to take vC up to U2.  Hoping for U3.  There are 2 vCs in Linked Mode, are there special considerations to take into account - I've never upgraded linked vCenters?  I can't break this system! 

0 Kudos
AmzSannegowda
Contributor
Contributor

0 Kudos
FJ1200
Enthusiast
Enthusiast

Thanks - Not seen that one - running it now.  They do seem to have a major issue with snapshot sprawl!

0 Kudos
FJ1200
Enthusiast
Enthusiast

There don't appear to be any snapshot issues.  I upped the ThreadStackSizeKb to 1024 as per the kb article and it did last a lot longer but it still fell over with the same error.  We're going to reboot it early this week, apparently this one has history!  We do have some datastores that are full, and all hosts in this cluster show the same errors in the logs - not enough space for journals.  Datastore space is being addressed, but this host does not have any vms residing on the full datastores, but they are mounted on the host.  Will that affect re-joining?

<update>

I'm going to update one vCenter in the morning and see how it goes, then the 2nd probably the following day.

0 Kudos
FJ1200
Enthusiast
Enthusiast

Anyone got any ideas?

I did discover yesterday that a 5.0U1 vCenter can manage 5.0U2 ESXi hosts, although Best Practice is still to have everything on the same level so looking at updating today.

VMware Product Interoperability Matrixes

Didn't know this existed.

0 Kudos