Duplicate UUIDs

[martin_zidu]

On AMI Bios you can use AMIDEWIN tools or DMIEDITOR tools... I have had problems on some customers... it's time consuming and cost a lot to change all bios UUID...

I have had this issue at industrial customers on industrial workstations (panel PC, and fanless computer, all AMI BIOS...)

Yes, you are right that those duplications are mainly on production PCs and terminals...because it seems to us, that manufacturer is sending machines with cloned BIOS...

We have also some set of tools from technician, which will provide us with a possibility to change UUID, but still it´s a "hardware layer" change and this is something, which is really not good idea to do...according to the fact, that sometimes its 24/7 production machines and we are not able to find suitable maintenance window to do a change.

We had also some office PCs affected, but this is caused due to the fact, that they have corrupted UUID (like empty FFFFFF-FFFF-FFF....) and therefore sometimes can be also an issue.

[bastiaenssenst]

If Veeam would use the MAC-address of a computer maybe in combination with BIOS UUID most of the problems would solved i guess

[shawmcbigdis]

I'm also having this problem with a few white box PC's that are used for a custom application here.

Anyone have any more info on a fix from Veeam on this?

I have a ticket open with support, but the are just having me remove the PC's from the protection group and manually load the agent over and over. It would seem if the R$D team know this doesn't work, then they need to tell support there is no fix right now, instead of telling people to jump through hoops to accomplish nothing.

[martin_zidu]

Good Afternoon Gentlemen,

procedure is as follows:

1) First of all, this procedure is working only with an U3a update

2) To ‘EnableSystemUuidFailover’ key should be added on each Veeam Agent host with non-unique ID.

This is the key enabling Veeam failover logic.

Name: EnableSystemUuidFailover
Path: HKLM\SOFTWARE\Veeam\
Type: DWORD
Value: 1

3) Then a unique ID should be assigned manually on each Agent station.
Another Veeam registry should be added on each host for it:

Name: SystemUuid
Path: HKLM\SOFTWARE\Veeam\
Type: REG_SZ
Value: {a custom UUID}

Custom UUID should correspond with the standard UUID format rules, brackets are required.

You can use some online generator for it (https://www.uuidgenerator.net/ ) or put some your own strings there.

4) Rescan protection group and you are deployed now

BTW: obviously it´s still an workaround and Veeam is suggesting to update BIOS UUID together with vendor since UUID duplication could lead not only into Veeam agent issues.

Hope this will help to some of you guys

We have tested and it´s working

point is, if this solution can be cover by group policy or similar...but thats different story how to handle those registries

Cheers

Martin

[bastiaenssenst]

I can confirm this workaround works.
Tested this on several machines.

[sockpuppet]

Have you been through the BIOS and checked for an option to change part number and serial number? I recently had an HP server with no UUID and this was because the serial number had been removed, once I put the serial number back in the BIOS, the UUID was created from the two fields and Veeam backup worked.

[ddayton]

I would just like to add that without using Sysprep after a clone has caused us many issues with active directory (AD), McAfee ePO, SMS and WSUS. You shouldn't clone without doing a Sysprep. It does not take much time to do, but good luck.

[TrevorBell]

@ Dima P

Was this supposed to be fixed in Update 4 ? Im sure i was told the fix would be in the release ?

23/01/2019 06:27:46 :: Unable to install backup agent: cannot connect to uk-xxxxxx Error: Known agent uk-xxxxxx have the same bios uuid 03000200-0400-0500-0006-000700080009 but different connection point uk-xxxxxxx

I would rather fix this issue with Veeam than any sort of reg hack if possible ( and yes buying physical servers with supermicro motherboards was not a good idea but these are CCTV Servers )

Will open a support case to investigate later today after you have commented.

Many Thanks

Trev

[Dima P.]

Hey Trev,

Thanks for sharing. Some fixes related to this issue are included in the Update 4, but to understand what's going on we still need the logs, so please open the case and share the case ID with us. Cheers!

[TrevorBell]

Will open a case and see if it can be resolved or not many thanks!

[arlosse]

Hello,

I've been facing a UUID duplication issue too.
We use the free version of Veeam agent for Windows, backup are sent to a Veeam repository (v 9.5 u3a).
At first, everything looked fine. But when I enabled Veeam Backup Copy job to send a backup copy off-site for our physical PCs, I had an error regarding duplicated UUID.
Only one of the concerned computers was listed in the backup copy job results (2 computers have the same UUID).
Digging a bit further, I discovered that both Veeam Agent Backup tasks were showing the same computer name in the backup results (but both computers do have a different name at operating system level).
I would not have been aware of that without the Backup copy job as no error or warning were shown on Veeam agent backups.
The registry key workaround worked, everything seems to work fine now.

However, Veeam support stated this cannot be considered as a definitive solution as it can cause issues in merge and restore operations.

Changing UUID at BIOS level is not an option. This are industrial computers that are driving production lines, I cannot take any risk with them.
Please, consider a descent way to deal with this issue in future releases.

You can check tickets 03425697 and 03446407 for more details.

[compdoc]

I have a server with a xeon cpu and a workstation with a pentium cpu, both have the same bios UUID because of their AMI bios. I managed to change the bios UUID in the workstation because it will boot Freedos, but then I had to delete all the disks and inventory of computers and add them back so the old UUID was forgotten. The server will not boot dos or Freedos, so it cannot be changed so far.

After googling the error I see others have had the same problem for years, and Veeam has done nothing about it. What a disaster.

[controlfreak]

I ran into this issue on agent version 3.0.1.1039 (so it is not resolved yet).

The fix suggested by martin_zidu still works. I created the reg keys on a system where the remote agent install was failing. Once the registry key and new UUID were configured, the agent installed successfully.

-Control

[Dima P.]

Control,

Any chance you have a case ID to share for a review? Thanks in advance!

[marko.grzan]

Does anyone have workaround like martin_zidu suggested for windows, but for linux machines as in my case i have linux VM-s with same UUID?

[Dima P.]

Hello Marko,

The proposed workaround works for all backup agents. To make sure it works for your issue - please raise a case and check with our support team. Thank you!

[nunciate]

Well, I just had this issue with a single server that has been backing up just fine for over a year now. Very odd indeed.
I have 37 Cisco UCS-E servers all in remote branch locations and I back them up nightly over the WAN. Everything works fine and then today I come in and see just one of them won't back up anymore.
I have been using the latest agent version for quite a long time now. I have changed nothing other than this was MS patch weekend so we did apply MS patches on Saturday.
After the reboot, this one server gave the biose UUID issue. I was able to perform the workaround from martin_zidu's post and that did fix the issue but this just seems very odd that it would suddenly come up.

[Dima P.]

Hello Alan,

Can you please raise a support case and share the case ID? Debug log analysis is required, so unfortunately I cannot comment your issue without the help of our support team. Thank you in advance!

[ViktorG]

Hello,

have the same problem that the BIOS uuid of several production machines is the same, Support - Case # 03974173.
Unfortunately all suggestions did not help with changing the BIOS uuid in the registry.
The company that is now getting a new production machine but has the same problem that the BIOS uuid is the same.
They are all Windows 10 computers.
Who has an idea what else could be done?


Thanks and greetings
Viktor

[Dima P.]

Hello and welcome to the community Viktor,

Please keep working with the support team since debug log analysis is required to identify the root cause. Cheers!

[g-rant001]

Well I have been testing this issue today with B&R version 10. I could not for the life of me to get the scan to work UNTIL I changed the system from the DNS name to IP. Come on man!! Now I will have to find a way to properly remove the system with the IP and then try and add it back with DNS.

[g-rant001]

Ok so my testing found that by creating a new protection group(test) and adding the system in question by the IP. Running a scan and then I selected the system in question and told B&R to uninstall all components on that system. Then I deleted the protection group. Finally the system was added to my Domain Computers group via the DNS name and everything is working now.

I find it a shame that you cannot cleanup the database so that you as the admin can tell Veeam that system x.x.x.x isn't there anymore.

[ROSA]

How do I fix this problem with VAL?
Case: #04725069

[PetrM]

Hello,

Let's wait for the reply from our support team. I'm sure they will be able to explain the behavior based on debug logs analysis.

Thanks!

[ROSA]

Both PCs had the same BIOS UUID.
Solved by support by:
stopping the veeamservice
deleting /var/lib/veeam/veeam_db.sqlite
creating /etc/veeamagentid with a new {UUID}
starting veeamservice

[AlexHeylin]

See also support cases 04833772 & 04426756 where a server running VBR is liable to change the BiosUuid any time it's power cycled. This makes VBR screw up with the CC and we have to send an active full to SP CC repo to start a new backup chain. It also causes duplication of the system in VSPC which then fires alerts and the old one has to be manually removed.

Veeam - this ID alone is not suitable for the use you're putting it to. There are numerous circumstances under which it's not unique, and/or not a constant.

[ejenner]

Just had this problem and goolged the error message and this thread came up.

My scenario is slightly different but just in case it helps somebody in the future I'll add the details. In my case it was just a small setup with Community Edition license.

I was using Veeam to duplicate an old server onto new hardware. Not sure if this is supported or not... but you can make it work.

I backed up the new server so I could create the Veeam Recovery Environment USB boot drive. Booted into the recovery environment and restored the system volume. This made it identical to the old server it was supposed to be a duplicate of.

Unfortunately in Veeam this had the effect of creating two windows agent computers with the same ID. Even though the computer names and IP addresses were different Veeam thought it was the same server. This was blocking me from restoring the old data volume from the old backup to the new duplicated hardware.

In my case with my scenario I restored my Veeam configuration from a configuration backup I took a few days ago before I started this migration process. By doing that I removed the new server from the configuration and Veeam was then working with just one server as it did before. Therefore no duplicates detected.

Surprisingly when I went to perform the restore of the data volume my backup from today was available as one of the increments. Even though it was from a point in time later than the configuration I restored. Not complaining.

This is Veeam V10. So the problem still seems to be there in one way or another.

[Dima P.]

Hello,

Thanks for the feedback!

Unfortunately in Veeam this had the effect of creating two windows agent computers with the same ID. Even though the computer names and IP addresses were different Veeam thought it was the same server. This was blocking me from restoring the old data volume from the old backup to the new duplicated hardware.

Good point, I'll discuss it with QA folks. Possibly we can add optional switch to forget older agent deployment as soon as new one become visible.

Surprisingly when I went to perform the restore of the data volume my backup from today was available as one of the increments. Even though it was from a point in time later than the configuration I restored. Not complaining.

Yup, we specifically tested this scenario for instant recovery case: when machine goes down you can power it on via instant recovery, use it temporarily, take backup when it's time to power on the physical host and restore latest state to a new physicals machine.

[ejenner]

Possibly over complicating it? What seems to be missing, i.e. your first instincts when this problem comes up is the ability to right-click and delete a managed agent from the console. I don't think it has to be as complicated as providing an override function.

[Dima P.]

EJ,

It's possible via UI in v11 - Removing computer from protection group.