Trouble with vmware plugin
Sylvain Donnet
I have installed the Vmware plugin, and it seems operational. But when I test it, on several VM with CBT enabled, I got the following error :
Fatal error: Traceback (most recent call last):
File "/usr/lib/bareos/plugins/BareosFdWrapper.py", line 42, in start_backup_file
return bareos_fd_plugin_object.start_backup_file(context, savepkt)
File "/usr/lib/bareos/plugins/vmware_plugin/BareosFdPluginVMware.py", line 160, in start_backup_file
if not self.vadp.get_vm_disk_cbt(context):
File "/usr/lib/bareos/plugins/vmware_plugin/BareosFdPluginVMware.py", line 887, in get_vm_disk_cbt
changeId=cbt_changeId)
File "/usr/lib/python2.7/dist-packages/pyVmomi/VmomiSupport.py", line 566, in <lambda>
self.f(*(self.args + (obj,) + args), **kwargs)
File "/usr/lib/python2.7/dist-packages/pyVmomi/VmomiSupport.py", line 375, in _InvokeMethod
return self._stub.InvokeMethod(self, info, args)
File "/usr/lib/python2.7/dist-packages/pyVmomi/SoapAdapter.py", line 1301, in InvokeMethod
raise obj # pylint: disable-msg=E0702
vim.fault.FileFault: (vim.fault.FileFault) {
dynamicType = <unset>,
dynamicProperty = (vmodl.DynamicProperty) [],
msg = u'Erreur provoqu\xe9e par le fichier /vmfs/volumes/19536d18-6870b949/vm124.ddo.net_1/vm124.ddo.net.vmdk',
faultCause = <unset>,
faultMessage = (vmodl.LocalizableMessage) [],
file = '/vmfs/volumes/19536d18-6870b949/vm124.ddo.net_1/vm124.ddo.net.vmdk'
}
"msg" could be translated by "error coming from the file ..."
The plugin setup is good, as it has found the vCenter and the VM.
Other VMs generate also this kind of message.
I have checked that these VM are really CBT enabled.
Any idea ?
BR
Sylvain
Stephan Dühr
unfortunately the VMware API does not give back the real cause for this error,
I've seen it before, and I think it was related to invalid CBT information.
Did you enable CBT with vmware_cbt_tool.py as described in
http://doc.bareos.org/master/html/bareos-manual-main-reference.html#x1-27800024.1.9
Could you please try to run it with --resetcbt, I think that fixed the problem
for me the last time I've seen it.
It may also be necesary to remove all existing snapshots before using vmware_cbt_tool.py
with --resetcbt
Regards,
Stephan
Stephan Dühr stepha...@bareos.com
Bareos GmbH & Co. KG Phone: +49 221-630693-90
http://www.bareos.com
Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646
Komplementär: Bareos Verwaltungs-GmbH
Geschäftsführer: S. Dühr, M. Außendorf,
J. Steffens, Philipp Storz, M. v. Wieringen
Sylvain Donnet
I have enabled CBT manually in my vCenter. So, now, I use the vmware_cbt_tool.py tool to resetcbt, or disablecbt/enablecbt on the VM, but this does not change the error. Of course, I have checked that there was no snapshot on my VM.
I also use the tool on a fresh VM, never CBTed. And backup also failed, with same kind of error message, on the corresponding disk.
Very very strange...
Sylvain Donnet
I have also a vCenter 5.5, which produces exactly the same error...
Sylvain Donnet
As it was not found in my Ubuntu 14.10, I downloaded it from here :
https://launchpad.net/ubuntu/+source/python-pyvmomi
It could be an obsolete or invalid pachage ?
Stephan Dühr
> And, I was tested that on my vCenter 4.1 (cluster of ESXi 4.1).
> I have also a vCenter 5.5, which produces exactly the same error...
>
I can't tell if 4.1 should work or not.
Regards
Stephan Dühr
installed, as there are none of them for Ubuntu 14.10 on download.bareos.org,
did you install the Debian 8 packages on Ubuntu 14.10?
The version of python-pyvmomi at https://launchpad.net/ubuntu/+source/python-pyvmomi
should be ok, the python-pyvmomi package that Debian 8 provides is version 5.5.0-2014.1.1-2
As of the error message, pyvmomi does seem to work fine, as it get's an error
message back from VMware.
Does the output of vmware_cbt_tool.py ... --info look similar to this:
INFO: VM stephand-test-tmp01 CBT supported: True
INFO: VM stephand-test-tmp01 CBT enabled: True
In the long run, you'd better use the one of the distributions listed at
https://www.bareos.org/en/faq/what-are-the-requirements-for-the-vmware-plugin.html
at least for one system to run a bareos filedaemon with the vmware-plugin.
Sylvain Donnet
Thanks for your returns.
I had installed Bareos packages with "Debian_8.0" settings.
The bareos-vmware-* installed are :
bareos-vmware-plugin
bareos-vmware-vix-disklib
The first one, bareos-vmware-plugin, has required a dependancy, not present in distrib : python-pyvmomi. This is why I have downloaded from launchpad.net website.
I also have installed bareos-vadp-dumper, but not set it up.
On the VMs tested, the "--info" works fine and returns True on "supported" and "enabled" settings.
Stephan Dühr
by the plugin once it could succesfully get the CBT information.
If you got that error, bareos-vadp-dumper will not have been run.
Searching for "vmware cbt Error caused by file" on google shows lots
of results, mostly pointing out that a CBT reset is required.
May be it helps if you exactly follow the instructions of
http://kb.vmware.com/kb/2139574
to reset CBT manually.
Stephan Dühr
> On 12/04/2015 11:13 AM, Sylvain Donnet wrote:
>> Hi Stephan,
>>
>> Thanks for your returns.
>>
>> I had installed Bareos packages with "Debian_8.0" settings.
>>
>> The bareos-vmware-* installed are :
>> bareos-vmware-plugin
>> bareos-vmware-vix-disklib
>>
>> The first one, bareos-vmware-plugin, has required a dependancy, not present in distrib : python-pyvmomi. This is why I have downloaded from launchpad.net website.
>>
>> I also have installed bareos-vadp-dumper, but not set it up.
>>
>> On the VMs tested, the "--info" works fine and returns True on "supported" and "enabled" settings.
>>
>
> There is nothing to be done to set up bareos-vadp-dumper, it is run
> by the plugin once it could succesfully get the CBT information.
> If you got that error, bareos-vadp-dumper will not have been run.
> Searching for "vmware cbt Error caused by file" on google shows lots
> of results, mostly pointing out that a CBT reset is required.
> May be it helps if you exactly follow the instructions of
> http://kb.vmware.com/kb/2139574
> to reset CBT manually.
>
> Regards
>
The Bareos VMware Backup Plugin does currently not work with NFS datastores.
For more details see the folowing FAQ article:
http://www.bareos.org/en/faq/does-the-bareos-vmware-backup-plugin-work-on-nfs-datastore.html
Sylvain Donnet
Is there a workaround, or will be a fix for the plugin ?
Leo Janssen
and finally drilled it down to an SSL issue with urllib3 that
was always complaining about insecure requests. Assume that warning
spoiled the result output from the cbt_tool.
After a rough and dirty deletion (not recommended) of the appropriate
section in file
/usr/lib/python2.7/site-packages/requests/packages/urllib3/connectionpool.py
the backup of my VMs runs smoothly.
Peter Allgeyer
Am Donnerstag, 17. Dezember 2015 17:36:51 UTC+1 schrieb Stephan Duehr:
> Are you using an NFS datastore?
>
> The Bareos VMware Backup Plugin does currently not work with NFS datastores.
>
> For more details see the folowing FAQ article:
> http://www.bareos.org/en/faq/does-the-bareos-vmware-backup-plugin-work-on-nfs-datastore.html
>
This post is a little older but still current. You're right that the use of the special "*" change ID is the cause of trouble with VMs on NFS datastores. If you migrate the datastore to VMDK, backup it with Bareos and migrate it back to NFS, bareos works with CBT enabled on NFS datastore, because the change ID was set initially by the first backup and CBT is supported by VMware on NFS datastores also. From my understanding it should also be possible to do a intiially full backup of the VM sitting on the NFS datastore which converts the thin layout to thick and sets the necessary change ID for later CBT backups. Do you think, that it would be possible to implement that in Bareos?
Best regards