Agent for Linux Recovery ISO - CIFS version

[Orangeyouglad]

Hello Everyone,

First time posting - sorry if this has been posted earlier or if there is a KB page somewhere.
Some very brief searching didn't show any hits for me.
If there are answers already please let me know.
If not I will raise a case.

Details first - question later

I am in early stages of testing / deploying Veeam Agent for Linux - downloaded the latest version just a few days ago.
Running into a couple of issues with SMB / CIFS network target for backup and recovery.
Issues appear to be center around CIFS version - specifically I have CIFS1 / NT1 disabled in my Samba configuration only allowing specific higher versions of CIFS / SMB protocol.

Environment:

• Agent Version : v2.0.1.665
  Source OS : Ubuntu 16.04
  Initial Kernel: 4.4.0-130
  initial cifs module ver: 2.08
  Target SMB Server OS : Ubuntu 16.04
  Agent for Linux recovery ISO Version : veeam-recovery-media-2.0.1.665_x86_64
  recover ISO Kernel: 4.4.36

Backup
During the initial backup job configuration I was unable to mount the SMB share.
Some research suggested the version of the cifs.ko module have been updated in later kernel versions and that might get me out of the weeds.

Subsequently I upgraded the Source OS kernel to latest Ubuntu HWE kernel ( aptitude install linux-generic-hwe-16.04 )
Prior to installing linux-generic-hwe-16.04 my Kernel version was 4.4.0-130 , after installing HWE my kernel version is 4.13.0-45 and cifs module has been upgraded to 2.09

After rebooting Source Server - I was able to successfully create the backup job and send the backup to disk over SMB / CIFS

Testing Recovery
Testing the recovery process involved booting the recover ISO in a new ESXi vm and attempting to mount the SMB share.
I received an error message indicating the recovery ISO was unable to mount the SMB share.

For testing purposes I adjusted my SMB server to enable NT1 SMB protocols and was able to restore the backup archive successfully.

After testing the recovery process I then reverted my SMB protocols and disabled NT1

Error Messages
The error message I am seeing when hitting the CIFS protocol mismatch is :
mount error(95): Operation not supported
Refer to the mount.cifs(8) man page.


Questions

Can the Agent for Linux be configured to specify the CIFS protocol to be used?
Can the recover ISO be configured to specify the CIFS protocol to be used?
Can the recovery ISO be upgraded to newer kernel version or cifs version?

Considering the massive issues with SMB1 wondering if I am missing something fundamental here.
https://blog.rapid7.com/2017/05/16/upda ... -for-them/
https://blogs.technet.microsoft.com/fil ... sing-smb1/


Thanks in advance,

Orange.

[PTide]

Hi and welcome to the community.

Can the Agent for Linux be configured to specify the CIFS protocol to be used?

Not in the current version, but you can mount the share manually via CLI using mount -t and specify version there.

Can the recover ISO be configured to specify the CIFS protocol to be used?

You can drop to the command line and use the same approach as described.

Can the recovery ISO be upgraded to newer kernel version or cifs version?

Yes, you can build a custom recovery media.

Thanks

[Orangeyouglad]

Thanks PTide

Can the Agent for Linux be configured to specify the CIFS protocol to be used?

Not in the current version, but you can mount the share manually via CLI using mount -t and specify version there.

So I haven't tested yet but guessing that some Pre-backup / Post-backup scripts to manually mount and dismount the share
and having the Linux Agent use a local file path

Something along the lines of
mount -t cifs -o credentials=/root/.cifs ,rw,soft,vers=xxx //myserverhere.com/cifs_share /mnt/cifs_share

where vers=xx is the required CIFS version eg 3.1/ 3.0 etc

Can the recovery ISO be upgraded to newer kernel version or cifs version?

Yes, you can build a custom recovery media.

Initial testing appears the kernel and modules are updated when the ISO is modified giving me the same kernel and modules as the server I am backing up.

Thanks for the quick reply.

Cheers.
Orange

[PTide]

So I haven't tested yet but guessing that some Pre-backup / Post-backup scripts to manually mount and dismount the share
and having the Linux Agent use a local file path

Yes, usage of pre/post-job scripts to mount/umount the network folder will work.

Thanks

[RPTE]

I had this same error and I resolved it by changing the SMB version on the NAS. Seems that it no longer works with SMB1