I'm still failing to get an image to reload after capture. It's
independent of ESXi version (4.1, 5.1, 5.5). It doesn't appear to be a
database issue (dropped the database and reloaded it from scratch and
reconfigured by hand). It doesn't seem to be a cygwin issue (tried both
32-bit and 64-bit). It doesn't seem to matter what version of Windows
(I've installed 8.1 and captured successfully and deployed a 7 instance
from another VCL and prepped it minimally and captured successfully). I
can create a new VM through vsphere client and attach the vmdk to it fine
and run it.
There are two functions that appear to be failing. One is during capture
(near the end just prior to reload):
|7575|61:61|image| ---- WARNING ----
|7575|61:61|image| 2014-08-08
18:15:16|7575|61:61|image|image.pm:process(235)|unable to retrieve size of
new revision: vmwarewin7-win8121-v0, old size will be used
|7575|61:61|image| ( 0) image.pm, process (line: 235)
|7575|61:61|image| (-1) vcld, make_new_child (line: 571)
|7575|61:61|image| (-2) vcld, main (line: 350)
and so it assigned the value of 1450 to database vcl.image.size.
The other is
|15513|8:8|reload| ---- WARNING ----
|15513|8:8|reload| 2014-08-13
16:19:01|15513|8:8|reload|VMware.pm:check_vmhost_disk_space(2393)|failed
to determine additional bytes required for the vmx and vmdk directories on
VM host node4
|15513|8:8|reload| ( 0) VMware.pm, check_vmhost_disk_space (line: 2393)
|15513|8:8|reload| (-1) VMware.pm, load (line: 475)
|15513|8:8|reload| (-2) new.pm, reload_image (line: 671)
|15513|8:8|reload| (-3) new.pm, process (line: 291)
|15513|8:8|reload| (-4) vcld, make_new_child (line: 571)
|15513|8:8|reload| (-5) vcld, main (line: 350)
When I add debug lines to VMware.pm it looks like the vmdk function is at
fault. When I change the vmdk function to just return 1 it still throws
the same error as above. The lines I add are:
if (!defined($vmx_additional_bytes_required)) {
notify($ERRORS{'WARNING'}, 0, "failed to determine the
additional bytes required for the vmx directories");
}
if (!defined($vmdk_additional_bytes_required)) {
notify($ERRORS{'WARNING'}, 0, "failed to determine the
additional bytes required for the vmdk directories");
}
Complete logs can be found in previous emails.
Thanks again,
jeff
From: Jeffrey Kirby <[email protected]>
To: [email protected],
Date: 08/11/2014 08:27 AM
Subject: Re: First Base image on new VCL won't load after capture
And this is not an ESXi 5 issue. This weekend I tried moving an ESXi 4.1
blade from the old, working VCL instance to the new one. Changes
necessary were local datastore names and ssh authorized_keys. I created a
VM from a OVA of a working guest I made on the old VCL, assigned it
correct MACs, and reran its cygwin-ssh-config and gen-node-key. It failed
with the same messages. Somewhere I have a boo boo in my configuration.
Thanks again,
jeff
From: Jeffrey Kirby <[email protected]>
To: [email protected],
Date: 08/08/2014 06:26 PM
Subject: Re: First Base image on new VCL won't load after capture
Attached is the log from a new windows 8.1 capture from start to finish
(failed reload).
Thanks!
jeff
From: Andy Kurth <[email protected]>
To: [email protected],
Date: 08/08/2014 02:48 PM
Subject: Re: First Base image on new VCL won't load after capture
I can't tell from the output if it's failing to determine vmx or vmdk
space. It would be helpful if you change VMware.pm line 2393 to try to
output the values it is checking, change it to:
notify($ERRORS{'WARNING'}, 0, "failed to determine additional bytes
required for the vmx and vmdk directories on VM host $vmhost_name, vmx:
$vmx_additional_bytes_required, vmdk: $vmdk_additional_bytes_required");
This will generate an ugly "failed to concatenate an undefined variable"
warning in vcld.log but at least we will be able to tell which value it is
failing to retrieve, $vmx_additional_bytes_required or
$vmdk_additional_bytes_required.
Also, please include the entire vcld.log output for the entire process.
Some information may have been retrieved earlier on in the execution which
could be helpful for debugging.
-Andy
On Fri, Aug 8, 2014 at 11:33 AM, Jeffrey Kirby <[email protected]>
wrote:
I've now also added an ESXi 5.1 host and it can't load to that, either,
same error.
Running vcld with the switch -d doesn't seem to add any debug info to the
log. Am I missing something?
From: Jeffrey Kirby <[email protected]>
To: [email protected],
Date: 08/06/2014 02:07 PM
Subject: First Base image on new VCL won't load after capture
I am trying to deploy a new VCL instance. I'm at the point where I have
captured a windows 8.1 image (64-bit, 64-bit cygwin; I altered Windows.pm
to correct the location of Sysnative as system32). The problem I'm
experiencing is that on redeploy at the end of the capture it errored out
with not being able to determine additional bytes required for the vmx and
vmdk directories on VM host. I can recreate the issue without capturing
by just requesting an image reload through Manage Computers > Computer
Utilities.
[attachment "marist_vcld.log.gz" deleted by Jeffrey Kirby/ADM/Marist]
