Video Screencast Help

Error opening the snapshot disks using given transport mode: Status 23.

Created: 21 May 2010 • Updated: 27 Jun 2010 | 7 comments
This issue has been solved. See solution.


We are using Netbackup 7 and are attempting to backup a vmware client via fiber when we receive the following error:
"Error opening the snapshot disks using given transport mode: Status 23."

The VMware guide states the following for this error:

"If the VMware backup host cannot access the VMware datastore by means of
the transfer type that was selected for the job, the backup fails.
Select a different transfer type and retry the backup."

The response is pretty obvious ... The transfer type can easily be changed from the policy, although we have it set to SAN for a reason.

Our Configuration:
Master server - Windows 2003
Media server / Host backup server - Windows 2008

VMware ESX 4.0

Both media and vmware host are in two different zones, although the shared snapshot disk is located on the SAN which both have access to.

Any recommendations? Is there a port that needs to be opened on my media server to allow the message to work?

Thank you.

Comments 7 CommentsJump to latest comment

Marek Kedzierski's picture

shared snapshot disk ?
Snapshot is stored on different VMFS than original disk ?

Ira Brock's picture

Also, we are not zoning the disk with the backup server as I want to make sure the backup server never sees the physical disk as a device where it could run the risk of messing something up.

Does Symantec require this to work? It does not mention it in their documentation if so. Come to think of it, short of mentioning how to install and configure the settings, it does not mention topology once in their documentation?

Anyone else run a configuration like this? .. Transfer type: SAN   - VMware backups?

Symantec support is a joke, I have already tried that route a number of times.

Will Restore's picture

definitely needs access; Netbackup Master does not. 
See Figures 1-1 and 1-2 in  Symantec NetBackup ™ for VMware Administrator's Guide

Will Restore -- where there is a Will there is a way

Ira Brock's picture

Thank you for the reply.

I agree. The netbackup host definitely needs SAN connectivity to the master. I do have it set this way already as well. This is needed for the basic media master communication.

I was finaly able to get a VMware tech in Symantec to take the case. He said "The VM-Ware documentation for Netbackup does not include everything. This was done intentionally as some of the set up is done from the VMware side, and if it is in our documentation, we must support it, and we do not."

Given this. He said that the SAN vmware host backup server needs to be zoned in the SAN fabric to see the shared HDD's the snapshots reside on. This was not given because it is not a good idea ... but it is the only way it will work. Originaly, Symantec said it just needed to see the ESX host in the zone, but now, they are saying it actually needs to see the physical LUN on the SAN to get read/write access to the HDD where the snapshots live.

I will be setting up a new zone that incorporates this and will let you guys know if it fixes the issue. Hopefully if it does, this can help many others as this question is all over the Internet without a proper resolution from Symantec, as they do not wish to get their hands dirty and include it in their VM-Ware guide. (They could of at least referenced a VMware backup guide made by VMware.)

Ira Brock's picture

To gather more logging information, perform the following:

1) On the backup host server open the Backup, Archive and Restore GUI.

2) Select FILE 
4) Go to the TROUBLESHOOTING tab 
5) Change GENERAL to 2 and VERBOSE to 5 
6) Close the GUI 
7) Browse to \\veritas\netbackup\logs 
8) Create a folders/directories named bpfis and bpbkar
9) Rerun failing policy(s). 
10) Check Logs

Also, not mentioned in the guide, in order to get more detailed logging information for the VM-ware backup, perform the following in the registery of the Host-Backup server:

Change the logging value to the decimal value of 7400 (put back to 0 once you have finished)

HKLM\SOFTWARE\Veritas\Netbackup\CurrentVersion\Config\   Create a key called "BACKUP" (all caps)
Create a dword value called "vmloglevel" and set it to 3. (3 - 6 is a valid entry.)

These keys will add more information to your logs regarding the vmware side of the backups.

Ira Brock's picture

Adding the host ESX machine to the same zone as your Host-Backup server does not help. You must zone in the physical disks on the SAN LUN so that your Host-Backup server can see them in the Disk Management snap-in. In order to do this, Add the LUN to a new (or existing) SAN zone within the SAN fabric. 

After doing so, tell the Host-Backup machine to scan for new devices in the device manager snap-in. Then Check your Disk Management Snap-in to verify the disks are showing up. They should show up as basic discs without drive letters, just as if the SAN was sharing them to the machine to be mapped. (Do NOT map them).

Re-run the job, and it should now work as the Host-Backup server has full access to the physical hard disks that the snapshot resides on.

** My issue was that:
A) The Veritias / Symantec  guide for configuring VM-ware did a good job at leaving the topology out.
B) Once I finally managed to get support, they told me to only zone the physical ESX machine and the Host-Backup server together.
C) Once I finally insisted to speak with a better tech, I was informed that I needed to share the physical Hard Disks of where the snapshot actually resides on with the Host-Backup Server. After doing so, the backup worked flawlessly. (15GB in under 10min from FC disk to LTO4 via 2/4 GB Fiber Channel.. Not to mention the new ability of being able to backup Virtual Machines a lot easier) 

*Edit: Here is the link to the Symantec NetBackup 7.0 for VMware Guide. -
Helpful up to the point of the topology in which the above should help with.