Solved

Anyone moved an M365 Repo from one drive to another on block storage?


Userlevel 7
Badge +22

Hi folks,

My wanderings in the wonderful continue. I need to move an M365 repo from a REFS drive (as far as I know this is not good, NTFS is the best practice) to a new drive. I have read on the forums that the Move-VBOEntityData powershell is NOT the way to go. Instead this is the procedure: 

1. Disable all backup jobs related to the repository you want to move.
2. Stop all the Veeam Backup for Microsoft Office 365 services before moving any data.
3. Copy the backup repository data to the backup repository location on the new repository location (using robocopy for example).
4. Start the Veeam Backup for Microsoft Office 365 services.
5. Add the new repository with the moved data.
6. Open backup job settings and specify to use the new repository

 

Any comments or advice from folks who have been through this. I would like to avoid any pain since I had a wisdom tooth pulled this week and have enough already :) 

 

cheers

 

icon

Best answer by Mildur 25 July 2023, 05:51

View original

13 comments

Userlevel 7
Badge +21

Those are definitely the steps.  We did this to move most of the ReFS drives we had to NTFS, but now moving to Object at some point.

I cannot recall though if the data on ReFS blew up when copying over so having enough space plus is a good idea.

Userlevel 7
Badge +22

I don’t think it would have any re-hydration since the jet databases would not be able to leverage spaceless fulls like the block cloning. I think that was one of the reasons why there was no point to put it on REFS but I also remember someone saying there also can be bad issues when on REFS, might have been one of the Vanguards at 11 systems. Or maybe @falkob. Either way S3 is the way to go as you said. 

Userlevel 7
Badge +6

I wasn’t aware REFS wasn’t recommended for VB365.  Appreciate that info….although I’m getting ready to move a repo to Object from an old existing server and then attach to a new existing server.

Userlevel 7
Badge +21

I don’t think it would have any re-hydration since the jet databases would not be able to leverage spaceless fulls like the block cloning. I think that was one of the reasons why there was no point to put it on REFS but I also remember someone saying there also can be bad issues when on REFS, might have been one of the Vanguards at 11 systems. Or maybe @falkob. Either way S3 is the way to go as you said. 

Yes, that is right too.  So, I guess then it just comes down to sizing of the database files you are moving and how long it will take.  Sorry brain is mush after VCC issues until 1:30AM this morning then back at Veeam stuff again today.  🤣

Userlevel 7
Badge +22

ReFS isn’t recommended exclusively because there’s a performance deficit if you have integrity streams enabled.

 

Details: 

I wouldn’t migrate to NTFS based on this I’d just disable the integrity streams option on ReFS instead 🙂

Userlevel 7
Badge +22

I don’t think it would have any re-hydration since the jet databases would not be able to leverage spaceless fulls like the block cloning. I think that was one of the reasons why there was no point to put it on REFS but I also remember someone saying there also can be bad issues when on REFS, might have been one of the Vanguards at 11 systems. Or maybe @falkob. Either way S3 is the way to go as you said. 

Yes, that is right too.  So, I guess then it just comes down to sizing of the database files you are moving and how long it will take.  Sorry brain is mush after VCC issues until 1:30AM this morning then back at Veeam stuff again today.  🤣

I know the feeling, I am zonked out on Tylenol but I am getting the feeling that my body is starting to adapt so I get only 2 hours relief instead of 6 but I can’t take two more pills until 6 so no choice but to try and stick with it. Mind starts to wander :( Same as sleep deprivation. :( 

Userlevel 7
Badge +22

ReFS isn’t recommended exclusively because there’s a performance deficit if you have integrity streams enabled.

 

Details: 

I wouldn’t migrate to NTFS based on this I’d just disable the integrity streams option on ReFS instead 🙂

Interesting to know. There is also a space issue involved so have to move it no matter what. Can’t use S3 yet either. :( 

I did not know about the integrity streams :) 

Userlevel 7
Badge +6

I just keep on learning…..thanks guys!

Userlevel 7
Badge +22

Folks any worries about this message “Changing backup repository may require performing a full backup. Previous restore points associated with this job will be unavailable”?

I mean that does not fill me with confidence to say the least. Since all the data is there should it not just see it and move on doing incrementals?

 

thanks

Userlevel 7
Badge +21

You should be ok to continue the backups. It is only a warning. 😜

Userlevel 7
Badge +12

Folks any worries about this message “Changing backup repository may require performing a full backup. Previous restore points associated with this job will be unavailable”?

I mean that does not fill me with confidence to say the least. Since all the data is there should it not just see it and move on doing incrementals?

 

thanks

Where do you saw that message?

First backup may be an full sync. Meaning the job needs to check if all items from M365 are in the new repository.
But it will run just an incremental download for the changed or new items.

All previous restore points will still be available.

 

Best,

Fabian

Userlevel 7
Badge +22

Folks any worries about this message “Changing backup repository may require performing a full backup. Previous restore points associated with this job will be unavailable”?

I mean that does not fill me with confidence to say the least. Since all the data is there should it not just see it and move on doing incrementals?

 

thanks

Where do you saw that message?

First backup may be an full sync. Meaning the job needs to check if all items from M365 are in the new repository.
But it will run just an incremental download for the changed or new items.

All previous restore points will still be available.

 

Best,

Fabian

That is what I am seeing. At first I thought “oh no, here we go, end of the show” :) but then I figured it would need to double check everything. I really want these folks to move to S3. 

Userlevel 7
Badge +10

Hi folks,

My wanderings in the wonderful continue. I need to move an M365 repo from a REFS drive (as far as I know this is not good, NTFS is the best practice) to a new drive. I have read on the forums that the Move-VBOEntityData powershell is NOT the way to go. Instead this is the procedure: 

1. Disable all backup jobs related to the repository you want to move.
2. Stop all the Veeam Backup for Microsoft Office 365 services before moving any data.
3. Copy the backup repository data to the backup repository location on the new repository location (using robocopy for example).
4. Start the Veeam Backup for Microsoft Office 365 services.
5. Add the new repository with the moved data.
6. Open backup job settings and specify to use the new repository

 

Any comments or advice from folks who have been through this. I would like to avoid any pain since I had a wisdom tooth pulled this week and have enough already :) 

 

cheers

 

@Geoff Burke

The method above requires the services to be stopped while the data is being copied/moved. That means for all the other VB365 organizations on that server, no backups will run.

You can use this method instead if you don’t want to shut down the services:

  1. Remove Backup Jobs for that organization from your VB365. Remember to capture all settings as these jobs will have to be recreated.
  2. Remove Backup Repository from your VB365.
  3. Copy/Move the files to the new location.
  4. If migrating to new server, add the same organization on the new VB365 server
  5. (Re)create Backup Repository pointing to new location. It will pick up the settings from the repository.
  6. (Re)create the Backup Jobs pointing to the new repository.

 

Comment