Moving VMs to a new virtual center.

[Greg69]

Hello,

As my vCenter server Appliance 5.0 has a problem and can't be updated (VMware support tell me this is a unknow problem ...), I have to make a new vCenter.
I will install vCenter 5.5 Server on Windows 2008R2.

As it will be recognize as a new vCenter server,
Is there any solutions for my Veeam server to continue Backup&Replication and not doing a full backup ?

I know that ID of my VMs/hosts will change, and I hope there may be a solution in the SQL DB of Veeam ... ?

Thanks for your help.

[veremin]

Hi Greg, once new vCenter is introduced, kindly, pause all existing jobs and open a ticket with our support team. As I've mentioned previously, there must be a script designed specifically for this use case. Thanks.

[cody.ault]

Hi Greg, once new vCenter is introduced, kindly, pause all existing jobs and open a ticket with our support team. As I've mentioned previously, there must be a script designed specifically for this use case. Thanks.

Support doesn't have a fix for the above issue. We have the following option http://www.veeam.com/kb1905 only if you are keeping the same vCenter database. If you make a new database, the only officially supported option is to add your new vCenter, then it's recommended to make a new job/clone the job and readd you VM's from the new vCenter to the job. You then need to let the jobs run new active fulls. Removing and readding VM's from the old jobs has caused some GUI related issues with restores in the past.

[veremin]

I think there is a slight misunderstanding, since I do know the guy from support who has written a script that addresses the said issue. So, if Greg can post his ticket number here, I will ask the script author to take the case over.

[marama]

Hi!
Any news here? I'm desperately waiting for the script to be able to migrate the VMs without needing to reseed the offsite repository.
My Case # 00663956 .
TIA

[Vitaliy S.]

Hi Marama, I will be able to check this on Wednesday and will let you know.

[cdl007]

We are working on the support tool to facilitate vCenter migration scenario, this should become available in the same time frame with v8. Thanks!

is it available ?

Christophe

[veremin]

No, currently it is not. The only thing I can tell you is that we're still working on it. No ETA at the moment, though. Thanks.

[aeccles]

No, currently it is not. The only thing I can tell you is that we're still working on it. No ETA at the moment, though. Thanks.

I'm very interested in this as well. I just came across this issue and I have about 2TB of data that will need to be reseeded over a 10mbps link The project is this weekend, so hopefully there has been some progress?

[foggy]

Then please contact technical support directly to get assistance in editing Veeam B&R database.

[aeccles]

Then please contact technical support directly to get assistance in editing Veeam B&R database.

Ok, I'll give it another shot. I just called in a few minutes ago and was told there was no way to do it. I'll tell them you sent me this time

[foggy]

If you post the case ID here, I will be able to ask the correct person to pick it up.

[aeccles]

If you post the case ID here, I will be able to ask the correct person to pick it up.

Case # is 00695458. I believe it's closed, so if I need to re-open it please let me know.
Thank you!

[fquinnAPT]

Hi Guys,

It looks like Vcentre needs to be rebuilt

I have a few options...I can rebuild and keep same DNS name and IP if I wish

Will VEEAM then simply carry on with the existing BU jobs once this new server is in place?

I know I can build a new server/appliance with new name and re point the jobs by re adding servers and mapping the backup in the job properties.

There is also this kb http://www.veeam.com/kb1905

I am just wondering what is the least painful ....

I have 2 existing jobs that I do not want to lose. One replication and one BU job

Thanks

[veremin]

When you register VMs under new vCetner, they will be assigned new MorefIDs and, thus, considered as new ones from backup server perspective. You can use mapping functionality without any issue, but be aware that new full backup will be created for those VMs. Thanks.

[fquinnAPT]

Thanks

This is not good for me ...I will have to try and fix the Vcentre server . I have been on with VMware and they have advised we build a new Vcentre server .....Surely there has to be a way of keeping these jobs in sync?

[veremin]

Otherwise, you can disable all existing jobs, re-register VMs under new vCenter, open a ticket with our support team and ask them to run custom script that updates MorefIDs inside product database. If your support engineer happens to be unaware of that, post your ticket number here, and I will ask an appropriate one to take over your case. Thanks.

[fquinnAPT]

Thanks again

I have a call open

If I have the same DNS name/IP is this going to be the case? Are the morefIDs on the Vcentre DB side?

[veremin]

If I have the same DNS name/IP is this going to be the case?

AFAIK, it should not be.

Are the morefIDs on the Vcentre DB side?

MorefIDs are present inside both vCenter and VB&R databases. You need to modify the latter one.

Thanks.

[aeccles]

Before you move them, install rvtools. I believe the last tab shows the IDs for each vm. Take a screenshot or export that info somehow. It will be helpful to the engineer that works on your case.

[fquinnAPT]

So to summarise

I am going to create new Vcentre appliance

Same DNS name and same IP address with new Vcentre DB. Will the VEEAM jobbs continue on ? Or create new Full backups ? Or do I need the script?

[veremin]

I must have been unclear in my latest reply.

You can preserve the same IP/DNS for new vCenter and VMs. Despite of that, re-registration of VMs will result in new VMs assigned to them. Without changes inside VB&R product database (custom support script), new full backup will be created for re-registered VMs.

Hope it makes sense.

[veremin]

Hi, Fintan,

May I ask you to post your ticket number here?

I've just found out that the script I was talking about is still in experimental stage. Only the script developer is currently using it on regular basis, meanwhile, others have nothing to do with that code. So, I don't want to put your engineer in an uncomfortable position of being at odds with what customers are being advised in the forums or even being uninformed of something.

Also, I think that my wording could have been better.

Anyway, provide me with the case number and I will try to ask the script developer to take a look at it.

Thanks.

[eringuet]

Hi,

Our company is currently migrating to a new domain. I also have to upgrade from vSphere 5 to 5.5. I have asked VMware for advice on how to move my vCenter to the new domain and if I should upgrade vCenter before or after changing the domain name.

Their answer was to do the following:

1. Move DB server to new domain
2. Create new VM for vCenter, join to new domain
3. Install vCenter 5.5 point to old DB.

This sounds straight forward enough but I wonder what will happen to Veeam?

Thanks

[foggy]

Etienne, provided that you've retained the original vCenter database and DNS name/IP, you should not get into any issues. If the name/IP changes, you would need to use hosts file on Veeam B&R console to point it to the right address.

[dcolpitts]

We are in the process of migrating our client sites from a full Windows vCenter 5.5 installation (with SQL Express on the VC) to VCSA 5.5. As a result of using SQL Express on the VC instance, we are unable to utilize the VMware Fling for migrating to VSCA. These sites are pretty basic in terms of vCenter settings (and usually less than 5 hosts and a 100 VMs), so tossing out the old vCenter database and starting fresh by reconnecting the existing hosts (and VMs) isn't a huge deal. Yesterday however, we hit the first site that relies on Veeam for their backups (the other sites we have converted were BackupExec only). Needless to say, Veeam puked afterwards, with every VM in the daily job failing with "Task failed Error: The object has already been deleted or has not been completely created".

Deleting the VMs from the job and re-adding them works - except for three issues, which I believe is because the VMs now have a different moref IDs. First, it appears CBT is reset, ignored or not recognized from the previous backup - the initial incremental backup is the same size as the weekly full active. Second, when I go to restore a VM from the job, each VM is listed twice - one with a few restore points, the other with 60 restore points, so I’m going end up with 120 restore points when their should only be 60 (this site maintains 60 restore points)… And third is VM retention deletion - we are using the default 14 day setting, so it concerns me that VBR may delete the previous 60 restore points in two weeks.

I can manually manage the three issues above at this first client site as we have the necessary physical resources to maintain double the backups right now, but this is not an option at some of my other sites, especially the sites that maintain daily restore points for a year and are running at 60 to 75% capacity in their backup repository.

Looking around, it seems like http://forums.veeam.com/powershell-f26/ ... t9301.html discusses updating the moref ID. Is this an option - run the PowerShell script that ThomasMc posted on Sun Dec 11, 2011 9:22 am? Does someone have a better idea, thought or theories?

I'm opened to suggestions.

dcc

[dcolpitts]

Looks like perhaps this may have worked for me... http://www.virtual-guy.com/?p=16

dcc

[foggy]

Dean, the observed behavior is totally expected and you're correct that it is caused by the moref ID's change.

[dcolpitts]

Looks like perhaps this may have worked for me... http://www.virtual-guy.com/?p=16

dcc

Ok - so http://www.virtual-guy.com/?p=16 did eventually function for me after I made a few modifications to the code (including formatting of quotes and stuff)... I don't know who's blog that was to thank him but he saved me a bunch of work over the next few weeks. To that end, here is my modified code based on virtual-guy's blog's code...

Creating the temporary SQL Table was ok so there were no changes here...

Code: Select all

USE [VeeamBackup]
GO
CREATE TABLE [dbo].[tengitafla](
[id] [int] IDENTITY(1,1) NOT NULL,
[Name] [varchar](500) NULL,
[UUID] [varchar](500) NULL,
[vmid] [varchar](50) NULL,
[bobj_id] [uniqueidentifier] NULL
) ON [PRIMARY]
GO

Next, was the PowerShell snippet - it wasn't happy due to some formatting issues and a few other odds and ends. Here is my updated PowerShell code:

Code: Select all

$sourceVI = Read-Host "Please enter the vCenter FQDN"
$creds = get-credential
add-pssnapin VMware* -ErrorAction SilentlyContinue -WarningAction SilentlyContinue | Out-Null
connect-viserver -server $sourceVI -Credential $creds
$report = @()
filter get-stuff
{
	$bleh = 0
	foreach ( $row in $_ )
	{
		$row = "" | Select ID, Name, UUID, vmid
		$row.ID = $bleh
		$row.Name = $_.Name
		$row.UUID = $_.config.uuid
		$row.vmid = $_.moref.value
		$row
		$bleh++
	}
}

$report = get-view -ViewType VirtualMachine |get-stuff
$report | Export-Csv -Delimiter ";" -NoTypeInformation -Path "$defaultviserver.csv"

$a = "$defaultviserver.csv"

(Get-Content $a) | Foreach-Object {$_ -replace '"', ""} | Select -Skip 1 | Foreach-Object {$_ + ";"} | Out-File $a

disconnect-viserver -server $sourceVI -Confirm:$false

Then the csv import into the temporary table had formatting issues too with quotes (make sure you adjust c:\temp\vcenter_fqdn.csv as required)

Code: Select all

delete
from [VeeamBackup].[dbo].[tengitafla]
bulk
insert tengitafla
from 'c:\temp\vcenter_fqdn.csv'
with
(
fieldterminator = ';',
rowterminator = '\n'
)
go

Next completing the temporary table.

Code: Select all

Update [VeeamBackup].[dbo].[tengitafla] SET [bobj_id] = BObjects.id
From BObjects
Where [VeeamBackup].[dbo].[tengitafla].Name = BObjects.display_name

Finding the ID of the new vCenter was good - The name and IP of my vCenter didn't change, it is the same as it was before, just in VCSA now, so I didn't have to add it to Veeam, and the AD credentials that were being used before with the Windows version still work with the VCSA.

Code: Select all

select id, name, type from hosts
where type = 1

And then when I looked at the data in [VeeamBackup].dbo.BObjects, because all my VMs had the correct vCenter ID, I changed the SQL query to not update the ID. And for some reason, virtual-guy was only updating [VeeamBackup].dbo.BObjects entries if the UUID was null and a VM... Personally, I think that should have been NOT null, and we have both VMs (type 1) and Templates (type 4) in our environment, so I changed the script to update anything that had a matching name.

Code: Select all

update BObjects
set [VeeamBackup].dbo.BObjects.object_id = [VeeamBackup].dbo.tengitafla.vmid
from [VeeamBackup].dbo.tengitafla, [VeeamBackup].dbo.BObjects
where [VeeamBackup].dbo.tengitafla.NAME = [VeeamBackup].dbo.BObjects.display_name

I also added a SQL Query to drop the temporary table as it was no longer required.

Code: Select all

USE [VeeamBackup]
GO
DROP TABLE [dbo].[tengitafla]
GO

Then as virtual-guy suggested, I opened Veeam, went to Backup Infrastucture and rescanned my vCenter server. Then I went to my daily job and removed my vCenter from it and saved the job. Finally I went back into the job and re-added my vCenter server back to the job and performed a recalculate to verify the Veeam saw my VMs.

The real test will be when tonight's forward incremental job kicks off, but I feel pretty confident now that I won't have any issues.

dcc

[bcconner]

Hi,

I had to split my network into 2 parts with the vSphere hosts going to a colo facility. I couldn't take the vCenter server to the colo facility (multi-purpose box. Good idea at the time. Sigh.)

Created a new instance of vCenter server, pointed it to the current hosts (2), and defined it to B&R. The old vCenter Server is unavailable but still defined.

How do I gracefully migrate to the new vc server? I can't remove it as it is referenced by current backups.

I can replicate with mapping to get that going again. I'm more concerned with backup/restore data, so I can access prior backups and not start over on 30 days' worth of data.

Thanks!

--Ben