Backing up to Azure Blob Storage

[jamesc23]

Hi,

I'm trying to set up a job to transfer files to blob storage in azure but I'm not too sure whether I've done it correctly as it's not working.

I've successfully set up cloud credentials as well as backup repository to connect to blob storage and I can see the folder created by Veeam in the blob container. I've set up the job in scale-out repository but I'm unsure of how to run this setup or turn it into a job.

Please kindly point me to the right direction. Thanks.

[HannesK]

Hello,
I moved your topic to the object storage forum.

please keep in mind that it is not a backup job setting. It' a scale out repository setting. It moves old ("sealed") backup chains to object storage.

Please see user guide how it works & how to configure

Best regards,
Hannes

[jamesc23]

Hi Hannes,

Thank you for your reply. I've followed the guide and had everything set up, but no files get uploaded to blob storage.

I had setup scale-out backup repository and point it to blob storage and set up the backup files that's older than 1 day to move to object storage, but nothing has happened.

What have I done wrong or what am I missing here?

[HannesK]

Hi,
I guess you missed the inactive / sealed backup chain part. Just give it time until GFS kicks in on your backup copy job.

Best regards,
Hannes

[waltcj]

Hi.

Am i right in thinking that if you were to use fuse and mount your blob to a linux vm which is in return added to a VBR deployment as a backup repository, you would be able to backup directly to blob, however, fuse would first bring back all the restore points to the linux server before they can be modified and thus the job would have to wait for this download process to finish before it can actually start transferring data.

Regards
Casper

[anthonyspiteri79]

Hey there Casper... taking a quick look at Fuse https://cloud.google.com/storage/docs/gcs-fuse you can certainly mount the Object Storage however this would not work with Cloud Tier as we talk directly to an Object Storage endpoint and keep indexes on what has been offloaded into the Capacity Tier.

You can potentially use the mounted Fuse file system to send backups directly to yes... however this will just send full files up to Object Storage and you will potentially suffer from performance issues.

fuse would first bring back all the restore points to the linux server before they can be modified and thus the job would have to wait for this download process to finish before it can actually start transferring data

What do you mean by this?

[HannesK]

Hello,
FUSE or any other "direct to S3" gateway technology will be extremely slow (not work) because we need to update metadata while creating the backup file. I'm not even talking about merges (depending on your backup mode). Merges will break everything.

Performance is also bad if you try to use any other 3rd party vendor hardware gateway.

Long story short: with the traditional Veeam backup file format it does not work. It needs to be the "optimized for blob storage" backup file format to get good results.

Best regards,
Hannes

[waltcj]

Hi anthonyspiteri79

From what i saw doing some testing i found that with a good enough internet connection the backups worked for small jobs. However, as soon as the Job reached its retention, it appeared that FUSE Downloaded all the files that had to be modified including the vbk, performed all the changes and reuploaded it back to blob. Even cloudberry appears to give horrible performance and if the job fails it cannot retry immediately as cloud berry is still uploading the previous files. From what i can see there is simply no way to backup to blob without using a sealed backup chain.

[HannesK]

that's what I meant with "not even talking about merges"