VMware Cloud Community
barracudasystem
Contributor
Contributor

P2V lost connection - but agent still logging

I was / am attempting a P2V of a Windows 2k3 R2 sp2 32bit server.

At around 15% the process errored and the Converter interface displayed a message, which I believe was "VMware Converter Agent  connection was lost".  Looking at the vmware-converter-worker log I can see that a log bundle was generated for this task (attached).  The log bundle shows the following error:

2013-01-23T11:47:39.983Z [03916 info 'Default'] [taskSpec,467] [task-3] [TaskMap] task-3:task-1
2013-01-23T13:18:38.289Z [03916 error 'Default'] [vdb,515] Commit failed: SQL_IOERR: disk I/O error
2013-01-23T13:18:38.804Z [03916 error 'Default'] [task,362] [LRO] Unexpected Exception: SQL_IOERR: disk I/O error
2013-01-23T13:18:38.804Z [03916 info 'Default'] [task,379] [task-3] -- ERROR -- Convert: vmodl.fault.SystemError
--> (vmodl.fault.SystemError) {
-->    dynamicType = <unset>,
-->    faultCause = (vmodl.MethodFault) null,
-->    reason = "SQL_IOERR: disk I/O error",
-->    msg = "",
--> }
2013-01-23T13:18:38.851Z [03916 info 'Default'] [diagnosticManager,267] Retrieved taskInfo for "converter.task.Task:task-3" mapping it to "converter.task.Task:task-3".
2013-01-23T13:18:38.851Z [03916 info 'Default'] [diagnosticManager,307] The task with id = "task-3" was found to be a "recent" task.
2013-01-23T13:18:38.851Z [03916 info 'Default'] [diagnosticManager,321] No existing log bundle found for task with id = "task-3". The task is still "recent" so a log bundle will now be generated for it.
2013-01-23T13:18:38.944Z [03916 info 'Default'] [diagnosticManager,790] Retrieving task related diagnostics for server task with id = "task-3".
2013-01-23T13:18:39.147Z [05616 error 'Default'] [vdb,515] Commit failed: SQL_ERROR: cannot commit - no transaction is active
2013-01-23T13:18:39.147Z [05616 warning 'Default'] Caught exception while processingFilterUpdate[0].ObjectUpdate[0]: SQL_ERROR: cannot commit - no transaction is active ...skipping -- class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > __thiscall Converter::Server::Job::JobProcessorImpl::ProcessUpdateSet(const class Vmomi::Core::PropertyCollector::UpdateSet &) ("d:/build/ob/bora-470252/bora/sysimage/lib/converter/server/job/jobProcessorImpl.cpp:1138")

However when we look at the vmware-converter-worker log we can see that the process appears to be running still.  Several hours later the log shows InitialFullSyncCloneTask @ 49%.

We have attempted to open Converter but the task / job does not appear in the interface.

Has anyone seen this before?  Is the P2V conversion still running and what happens when it completes if we can't access the task / job from the Converter interface?

Any help would be greatfully appreciated.

0 Kudos
3 Replies
patanassov
VMware Employee
VMware Employee

Converter server uses sqlite as db engine. Some error has occurred in it. Restarting "VMware Converter Standalone Server" service (don't touch worker and agent services!) should fix that. You may want for the conversion to finish before that to be on the safe side (monitor the worker's log, keep in mind there is reconfiguration after the cloning reaches 100%)

HTH

Plamen

barracudasystem
Contributor
Contributor

Thanks patanassov,

It's currently at 58% so I'll leave overnight and attempt the Server service restart once at 100% and post-processing is complete.

0 Kudos
barracudasystem
Contributor
Contributor

After the process finally finished processing we attempted the restart but the tasks/jobs didn't appear.

I think in our case there may have been some post processing errors aswell.

Anyway we reconfigured the disk configuration (split the source data drives accross several datastores on the VM host) and this error did not return.

0 Kudos