Cant not upload image to vmware bankend

[zehua]

A image is uploading to glance by the following commend:

glance --debug image-create --name cirros-0.3.4-x64 --container-format bare --disk-format vmdk --property vmware_disktype="sparse" --property vmware_adaptertype="ide" < cirros-0.3.4-x86_64.vmdk

In glance-api, debug a Traceback like this:

'Communication error sending http PUT requestto the url /folder/openstack_glance/e65f0d49-48e8-4df2-a124-47caf6367935%3FdcPath%3DDatacenter%26dsName%3Ddatastore236.\nGot IOError [Errno 32] Broken pipe'

[zehua (zehua)]

The problem code:
 url = loc.https_url
        try:
            response = session.put(url, data=data, headers=headers)
        except IOError as e:
            # TODO(sigmavirus24): Figure out what the new exception type would
            # be in requests.
            # When a session is not authenticated, the socket is closed by
            # the server after sending the response. http_client has an open
            # issue with https that raises Broken Pipe
            # error instead of returning the response.
            # See http://bugs.python.org/issue16062. Here, we log the error
            # and continue to look into the response.
            msg = _LE('Communication error sending http %(method)s request '
                      'to the url %(url)s.\n'
                      'Got IOError %(e)s') % {'method': 'PUT',
                                              'url': url,
                                              'e': e}

[Sabari Murugesan (smurugesan)]

Are you consistently hitting this issue ?

Can you check if the datastore you specified is not read-only ?

I think we need the corresponding vpxd.log when you do this operation. Can you provide that ? Also, additional information such as if the Datacenter or datastore are under inventory folders i.e in the Host's and Cluster's view of vCenter.

[zehua (zehua)]

I‘sure that datastore is not read-only. Where does the vpxd.log reside？

[Ian Cordasco (icordasc)]

This has been incomplete for almost an entire year. Closing as it appears to be a configuration issue and not a bug.