Meesed up Veeam B&R db after attemp to remove old vCenter se

[Marten_med_e]

[ID# 00759345] Remove old vCenter server from B&R server

It looks like I need to move to a new installation after a failed attempt to remove our old vCenter server from B&R, as we (me and Veeam support) tries to clean up the left overs, new errors pops up, mostly "Error: Object '94c09aa9-cd45-4ef2-9c05-306520150659' was not found" with different objects every time we remove one.
Right now a newly created job is failing due to a ESX 4.1 host that was deleted back in 2012 "was not found", but some jobs are working and some not.

But I have partly my self to blame for this as I kind of didn't take a db backup before running the ps cmdlet provided by support, and it looks like I had missed to delete some VM's from some "Excluded VM's" window in one or more jobs.

But I can make a new install of Veeam and a new db and maybe restore one of the backup configurations backups I have, but will that only make more of a mess? And what will happen with the backups that has been working, will the new jobs find the last backup time and continue or will it revert to date of restored configuration?

So what would be the best way to do it and a new installation after all those years and updates of Veeam don't seem as a bad idé.

1. New server with an new Veeam B&R installation
2. New db on SQL server
3. Restore Configuration backup before I did run the ps cmdlet
4. Import backup disks
5. Map recreated jobs to imported backup disks

or

1. New server with an new Veeam B&R installation
2. New db on SQL server
3. Import backup disks
4. Recreate "Guest OS Credentials"
5. Create new jobs and map recreated jobs to imported backup disks
6. Recreate "Cloud Credentials"


Cheers

Mårten

[Marten_med_e]

Forgot, for all Backup Copy Job that I have tried to recreate with GFS is failing with "Unable to map backup..." error.
Our cloud backup jobs is not using GFS so they are working, if the backup job they depends on is working.

Cheers

[veremin]

Forgot, for all Backup Copy Job that I have tried to recreate with GFS is failing with "Unable to map backup..." error.

It's expected behaviour. Currently, it's not possible to map a backup copy job to a chain containing GFS restore point. So, it might be worth waiting for patch 2 that will fix that issue. Thanks.

[Vitaliy S.]

So what would be the best way to do it and a new installation after all those years and updates of Veeam don't seem as a bad idé.

I would recommend using the first variant, and please note that you will not need to import backup files, as Veeam backup console will trigger a rescan procedure against all repositories. Given that all metadata files are in-place, you will see all backup files in the console, should be able to run the jobs.

If rescan does not find all backup files, then import them manually and then map your backup jobs to these files.

[Marten_med_e]

I guess so, I can manually import the backup but they never populate the repository and I can't map my new job on the new server to those files. Some do work and some don't on the same repository.
If I do a new job using the same name as it did on the old server, don't try to map it because I can't see it, then the job creates a new folder with the same name but a "_1" at the end, so the logic to not over write existing backup (?) works.

One that works on the new server but not on the old after first attempt from support to fix the errors


One that don't works, BUT it did work on the old server but not found in the new


I made a new support case as this is the "new" server and the "old" one is still in progress to see what went wrong when I was removing the old vCenter server.

Cheers

[Vitaliy S.]

If I do a new job using the same name as it did on the old server, don't try to map it because I can't see it, then the job creates a new folder with the same name but a "_1" at the end, so the logic to not over write existing backup (?) works.

Yes, that's expected. Your backup server is creating a new backup chain, as it has detected backup files with similar name created already on the repository.

I guess so, I can manually import the backup but they never populate the repository and I can't map my new job on the new server to those files. Some do work and some don't on the same repository.

Do you try to map backup jobs to these files or backup copy jobs?

[Marten_med_e]

Backup jobs, the Copy backup ones I have to wait to patch 2 or do the steps provided by the support tech, as some of them has GFS enabled.

The differences in naming of the vib files "Backup Job EQL-DS-082015-02-13T184255.vib" vs "1-Backup Job EQL-DS-082015-02-18T142338.vib" is because we (support tech and I) recreated the jobs on the old server as part of our attempt to clean up, all mapping did work at that time, but some jobs kept on failing with missing objects.

If we take this job as an example "1-Backup Job EQL-DS-08" it didn't work after I tried to remove the old vCenter server but when we did recreated it on the old server it could map to the files on the repository (separate Windows server) but it didn't run as it failed on a VM, but now when job is created on the new server with the same settings it works and I could map to the old files and all restore points are present.

But if we take this job"Backup Job NIM-DS-04" it didn't work after I tried to remove the old vCenter server but when we did recreated it on the old server it could map to the files on the repository (separate Windows server) and did run without failing, but when job is created on the new server with the same settings, I can't map to the files on the repository (witch is the same) it never populates when I do a rescan of the repository.

Cheers

[Vitaliy S.]

But if we take this job"Backup Job NIM-DS-04" it didn't work after I tried to remove the old vCenter server but when we did recreated it on the old server it could map to the files on the repository (separate Windows server) and did run without failing, but when job is created on the new server with the same settings, I can't map to the files on the repository (witch is the same) it never populates when I do a rescan of the repository.

Ok, but have you tried not to rescan the repository, but manually import the entire backup chain and then use mapping option?

[Marten_med_e]

Trying that right now, but are on the way home and using RDP from my phone, it works, but... Will see if I get the time to continuing when I'm get home.

Cheers

[Marten_med_e]

Ok, but have you tried not to rescan the repository, but manually import the entire backup chain and then use mapping option?

Got some 5 minutes before trying to get the kids to bed...

So I can import the backups, no errors and all restore points present


But still not map to it, it's not there


Veeam support tech is taking over the server tomorrow to see if we can figure out what's going on. Will try to move a backup from on repository to another to see if it gets discovered.

Cheers

Guess who is going to fall to sleep first, me or the kids

[Marten_med_e]

If I do a rescan of the repository after the import was done, I get "skipped 11"


But there are 25 folders on the disk all with both a vbk and a vbm file present, and if we look at this visible in the "mapping windows" is 8 jobs, so hmmm....
Guess I have to wait for tomorrow, is it OK to post parts from the log files on the forum?

Cheers

[Gostev]

No, please post the support case ID that contains said log files instead.

[Marten_med_e]

After a short nights sleep, I am starting to see a pattern here, I think.
After I moved a backup to a new repository I got an error when trying to import the backup from the vbm file "Unable to find storage Id = 7ced..... for link set", but if I choose to import from the vbk file the import works, but "only" 11 out of 14 restore points, better than nothing, but...

It does not show up in the mapping window, and then I did search around the web a little and found a forum post that kind of explain some of it http://forums.veeam.com/powershell-f26/ ... 24467.html

So if I guest right I got to get my hands on the script.

Support team has a script to recreate .VBM file. So, open a ticket and request it. Thanks.

Cheers

[Marten_med_e]

Now things are happening, I got a "Magic Veeam Tool" from my support tech, that makes it possible to edit vbm files....

So I open in notepad the vbm from a backup that don't work, I then compare every entry FileName="Backup Job STOVM092015-02-14T120852.vib" with what is really in the catalog, for this job one inc. restore point was missing.
The tool I guess compare the vbm and the vbk, anyway it showed that the file was missing from the vbk, so I just deleted the restore points made after the missing one and saved the vbm file with the same name.
Did a rescan of the repository and the backup was added, Yatta, this time I could map to the backup and start the backup job with out any errors.

So now I will try another job to see if it can be fixed this simple too.

Cheers