Discussions related to using object storage as a backup target.
Post Reply
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

copying data to IBM COS

Post by medacek123 »

Hello,
We need to find a way to copy all data from Bare-Metal LHBR to IBM COS.
Something about our infrastructure. It is approximately 200 TB of data for 50 VMs hosted in SOBR, created from 3 FS's as a performance tier. VBR server itself connected to IBM COS as a gateway. One 10Gbit network interface active
The copy option is not possible because I see an error about the non per-machine backup. Export is only possible for selected snapshots.

I think the best thing would be to add a new capacity and use the copy option. But I assume that if I choose to copy all data, it will copy all data in one step and use the maximum network speed. Every day there are running backups of PROD servers which takes 5-6 hours + every 3 hours app-awareness backup and restore to emergency copy. And that is the problem because if SOBR offload is active, the restore slows down significantly and cannot be completed within 3 hours. Any idea what other way to use or how to limit this SOBR offload? (maximum transfer speed is 300-400MB, so 8+ days if we stop all the other backups/restores)
Mildur
Product Manager
Posts: 10278
Liked: 2746 times
Joined: May 13, 2017 4:51 pm
Full Name: Fabian K.
Location: Switzerland
Contact:

Re: copying data to IBM COS

Post by Mildur »

Hello OC

You can use traffic throttling or <<<Offload Window>>> to control the offload behavior.

Best,
Fabian
Product Management Analyst @ Veeam Software
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

Hello, I got an idea. But do not know how it is working in detail.

For backups and restores can be remain, as gateway to bucket, VBR server itself.
For copying data to the new bucket can be used direct access or other gateway. What I understood from helpcenter in case of offloading data from LHBR for direct access, is used as gateway, LHBR itself. Do you think it will interfere somehow with VBR server, that data will flow through .

Another question is what ports must be opened for such offloading?
Mildur
Product Manager
Posts: 10278
Liked: 2746 times
Joined: May 13, 2017 4:51 pm
Full Name: Fabian K.
Location: Switzerland
Contact:

Re: copying data to IBM COS

Post by Mildur » 1 person likes this post

Hi OC
For backups and restores can be remain, as gateway to bucket, VBR server itself.
For copying data to the new bucket can be used direct access or other gateway.
The setting is configured in the bucket repository settings in Veeam Backup & Replication. It can be either direct or through a Gateway server and it affects backup, offload and restore sessions.

For offloading with a direct connection, your hardened repository will offload the data directly to the bucket. If you don't like that, you have to specify a gateway server.

The required ports are documented in our help center.
For S3-compatible storage, it typically requires Port 443 from the source performance tier extent (or gateway server) to the object storage target. You can also refer to the IBM documentation for IBM COS to find out which port they use to provide access to the bucket; I assume it is the default port 443.

The offload window can be used to limit the offloading to specific hours, and also the internet network traffic rule to throttle the throughput from any Veeam component to the IBM COS cloud service. You can configure the network traffic rule to only apply to backup/offload traffic by enabling the "Never throttle any restore activity" checkbox.

Best,
Fabian
Product Management Analyst @ Veeam Software
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

Hello Fabian,
thank you for confirmation. I added IBM COS as capacity tier to current SOBR, where are added all 3 FS's from LHBR as performance and changed to direct access for this IBM COS bucket. Now is offload of LHBR data running to IBM COS.
Once offload will be completed, we need to decommission LHBR. How can be removed this performance tier(LHBR FS's) or capacity tier(IBM COS) , for SOBR please, to be still able to work with copied to COS?
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

I did some test. I created local Win drive as a performance tier and COS bucket as capacity. Finished VM backup ,which was stored into performance tier and then it was offloaded to COS. Then I removed capacity tier from SOBR. After that I used import backups for COS repository, which was used before in SOBR and after completion of rescan , there appeared under Capacity tier(imported) data which contained before capacity tier in SOBR.

Do you know please other possibility how to do it?
Mildur
Product Manager
Posts: 10278
Liked: 2746 times
Joined: May 13, 2017 4:51 pm
Full Name: Fabian K.
Location: Switzerland
Contact:

Re: copying data to IBM COS

Post by Mildur » 1 person likes this post

Hi OC

If you want to remove your SOBR performance tier extents and only keep the IBM COS bucket, then thats the only way.
But please note, the backup server will not apply retention in that scenario. Backup on Capacity Tier are only removed by our background retention job if SOBR is still existent and offload sessions are running regularly.
If you reconnect your bucket used as a capacity tier, backups are imported and are not removed by the background retention job.

Best,
Fabian
Product Management Analyst @ Veeam Software
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

Thank you again Fabian, for your fast reply.

It is not a problem that will be not applied retention, because customer doesn't know if he wants original retention, which was used during backup or if he wants some longer retention as 1 year.

Regarding import backups I found documentation below:
https://helpcenter.veeam.com/docs/backu ... ml?ver=120
If I understood correctly, there is not need to do import backups and it is enough to rescan current COS repo? Because this COS repo is still added in VBR, only was removed from SOBR.
veremin
Product Manager
Posts: 20673
Liked: 2379 times
Joined: Oct 26, 2012 3:28 pm
Full Name: Vladimir Eremin
Contact:

Re: copying data to IBM COS

Post by veremin »

You’re correct—if this repository was previously used as a performance extent, then you simply need to rescan it. However, if the repository was used as a capacity extent, you’ll need to run the backup import procedure from it.
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

Hello Vladimir , thanks for your confirmation. Offload from local LHBR performance tier completed and all the data are copied now. I am going to start import and rescan.

But I do not understand one thing.

What I understood, in case of COS can't be used synthetic full backup, because of XFS missing, so I do not understand how is possible, that data stored in capacity tier size is only 75,1TB.
Original size without XFS savings is 139,6TB. In performance tier it was 82,2TB

So it means that for copying to COS was used some better dedup or compression or how it is possible?
medacek123
Novice
Posts: 7
Liked: never
Joined: Mar 12, 2025 2:05 pm
Full Name: OC
Contact:

Re: copying data to IBM COS

Post by medacek123 »

all the data are copied from performance tier(FS1-FS3) to IBM-COS

current SOBR copy utilization
name Type Capacity Free Used space
LHBR FS1 Hardened 54,8TB 33,4TB 39,1TB
LHBR FS2 Hardened 54,8TB 34,0TB 45,3TB
LHBR FS3 Hardened 54,8TB 14,8TB 55,2TB
IBM-COS IBM COS N/A N/A 75,1TB
Post Reply

Who is online

Users browsing this forum: Google [Bot] and 19 guests