HPE StoreOnce Catalyst Copy - Connection modes

[rbienvault]

Hello,

We have a case where the veeam infrastructure we have two separated infrastructure managed by the same backup server.
The proxies talk to the storeonces through FC.

When doing backup copy, it takes a long time to replicate all the vms on the replica repositories.
Catalyst copy feature seems to answer the issue and copy the repository but we cannot use it on our actual configuration as both FC network are autonomous and the catalyst copy use the same connection as it is registered on the backup server.

Is it possible to have a secondary connection method for the repositories, for example :
- for backup use that connexion (FC or network) and fill gateway server, etc
- if different for replication use network connexion and fill the gateway server, etc

Thank you !

[HannesK]

Hello,
the connection is not the problem. It's the StoreOnce systems (they should not be used as primary backup storage). If the copy is too slow, then CatalystCopy (or a whole re-design) is the only answer.

Catalyst Copy copies directly. A Veeam backup copy job de-hydrates the data.

Changing from FC to IP should also help, but it's not a silver bullet.

Best regards,
Hannes

[rbienvault]

Hello,

Yes, I agree with you, it is why I want to use that feature.
But I need to replicate the data using Catalyst Copy through network accross the datacenter and not through FC. On the same site, they are mounted through FC.

There is one limitation for the catalyst copy :
" Check that repositories between which you plan to copy data have a direct connection to each other. This is required because Veeam Backup & Replication uses the HPE StoreOnce Catalyst Copy technology to copy backup files.

This direct connection must be of the same type as the connection that you select when adding the target HPE StoreOnce. For example, if you connected the target HPE StoreOnce repository over Fibre Channel, you must connect the source HPE StoreOnce to the target HPE StoreOnce over Fibre Channel."

My suggestion is to have a secondary connection mode that can be use for the replication so the 2 storeonces can talk to each other.

Regards,
Romain

[evilaedmin]

@rbienvault

Will your environment allow you to copy the data as "file" objects (outside of Veeam) using the StoreOnce Catalyst Copy Utility? I would like to know if Veeam v10 has removed this limitation that exists in version 9.5:

https://community.hpe.com/t5/storeonce- ... zLE8i05QqI

the catalyst copy is using a field that has to be marked as completed.
Veeam does not modify that field, hecne the tool does not find any jobs to copy.

Catalyst Copy Utility manual

The StoreOnce Catalyst Copy Utility is a helper utility tool. It can be used to copy backup items toalternate StoreOnce Appliances for safekeeping, delete Backups that are obsolete or Orphaned,synchronize backup copies between a Primary Backup target and a Diaster Recovery site.

[HannesK]

Hello,
@rbienvault : As for the near future, your feature request will not be available. But take it as noted and we will check whether it is technically possible or whether there are other limitations, too.

@evilaedmin With V10 we have built-in catalyst copy support. We do not support external managed copies of our data as we cannot check whether the 3rd party tool did it correct (but we will be blamed if something went wrong).

Best regards,
Hannes

[foggy]

Hi Romain, if StoreOnce devices are connected via network, you can create a separate repository on the second StoreOnce via IP, point it to another folder (to avoid duplicates in UI/database) and use it as a target for Catalyst Copy jobs.

[rbienvault]

Hello,
Thank you for the response foggy.
Source will be still connected in FC, it will not connect to the target ? Or I missed something ?

[foggy]

If the arrays can talk to each other via Ethernet, Catalyst Copy will use that path. Veeam B&R can still connect to the source one via FC for its tasks.

[rbienvault]

Ok,
I will try that !
Thank you

[CLDonohoe]

@rbienvault can you confirm whether or not this solved your problem?

[rbienvault]

Hello,
Sorry for the extreme long delay.
I tested today and this has resolve my issue.

Thank you for your help !

[vcheca]

Finally, if the storeonces are on the same ip network, even if Veeam is not aware that this connectivity exists because they have been configured by FC, could the catalyst copy jobs be performed over the ip network? Is this supported by Veeam? How the storeonces are capable to do this replication?

[foggy]

Hi Victor, please review the Catalyst Copy job requirements:

This direct connection must be of the same type as the connection that you select when adding the target HPE StoreOnce. For example, if you connected the target HPE StoreOnce repository over Fibre Channel, you must connect the source HPE StoreOnce to the target HPE StoreOnce over Fibre Channel.

You need to add the target array using the same protocol you're going to use for the copy between them.

[vcheca]

I just found in the documentation the next:

The replication protocol is defined by the target backup repository. If CCoE is desired, make sure to define the destination repository as CoE and
not CoFC. Please note that it is supported to have the source repository accessed via CoFC, and the destination via CoE. If the destination needs
to be accessed via CoFC for performing restore operations, but the replication must use CCoE, then it is necessary to implement the following
workaround:
1. The production Veeam instance defines the target repository via CoE.
2. Install a new Veeam server to be used for restoring the destination replica via CoFC. The new Veeam server can be a VM.
3. On the new Veeam server, create a backup repository that points to the replicated copy via CoFC, which requires a server with FC
connectivity. Both the production server and the new Veeam server can use the same proxy/gateway, so you may not need to install a new
physical server just for restoring via CoFC.
4. Rescan the backup repository to import the restore point list to GUI.
5. The new Veeam server can be used for making a restore via CoFC from the Catalyst Copy replication destination.

This is a bit different..

Anyway, What is a way to replicate backups between two storeonces when they are configured with FC to their respective gateways and there is no FC connection between them?

[foggy]

That's actually doesn't contradict with what I've quoted - you need to add the target repository using the same protocol you want to be used between the arrays.

Anyway, What is a way to replicate backups between two storeonces when they are configured with FC to their respective gateways and there is no FC connection between them?

If there's no connection between the arrays, you need to add the target one to Veeam B&R via IP.

[rbienvault]

Yes, I confirm and it is working smoothly.

Veeam is only ordering the storeonce to create a replication task for a file.
The connection is established at the storeonce level without using the VEEAM Infrastructure to replicate the object on the catalyst.

Since the upgrade to VEEAM 11, we have issues with the catalyst task not launching (it very random), they are on a specific schedule.
I have not yet opened a case as it is stated that there is an know issue for the backup copy job. Do you want me to create one ?
Note : The issue is occuring almost every 2/3 days and I cannot stop / restart the veeam services as often.

[HannesK]

I don't know which known issue you mean, but if you like that someone verifies / helps you with that issue, then support always is the way to go (as stated when you click "post reply")

[foggy]

Hi Romain, yes, please open a case for a closer look as I'm also not aware of the issue you're mentioning. Thanks!

[rbienvault]

Hello,

I was talking about this know issue on veeam 11:

#8: Jobs don't start according to the schedule
Symptoms: Jobs (usually Backup Copy) do not start according to the schedule, get stuck in the Stopping state, or demonstrate other abnormal behavior.
Cause: Watchdog taking a vacation. Watchdog is a service thread which kills job sessions which did not have a chance to finish properly due to some external event. In v11, under certain circumstances Watchdog can terminate abnormally, resulting in the following unfinished job sessions hanging forever.
Status: Fixed in P20210507 or later.If you run into this situation, restart the Veeam Backup Service.

However, we did applied P20210507 and we still have the same issue.
FYI, case is progessing, and here is the reference : 04805018

Thank you !

[foggy]

Thanks for opening a case. Looks like you're not the only one seeing this issue - let's see what our engineers come up will after the case escalation.

[Hirosh]

Hi everyone,

is the Above case with the reference : 04805018 resolved? can anyone confirm this? or the issue still remains on V11?


best regards,
Hirosh.

[Hirosh]

That's actually doesn't contradict with what I've quoted - you need to add the target repository using the same protocol you want to be used between the arrays.


If there's no connection between the arrays, you need to add the target one to Veeam B&R via IP.

may i know why is it so? im just a little confused what is the logic behind the "direct connection must be of the same type as the connection that you select when adding the target HPE StoreOnce". i believe this is a big limitation. could anyone share some light on this?

[foggy]

The address and connection type used to add the target array to Veeam B&R is passed over to Catalyst for further processing, Veeam B&R doesn't know anything about other addresses or connection types the array may have nor is Catalyst itself able to detect those.

[foggy]

is the Above case with the reference : 04805018 resolved? can anyone confirm this? or the issue still remains on V11?

The issue discussed in the mentioned case was meant to be addressed in v11a.

[Hirosh]

The address and connection type used to add the target array to Veeam B&R is passed over to Catalyst for further processing, Veeam B&R doesn't know anything about other addresses or connection types the array may have nor is Catalyst itself able to detect those.

thanks @foggy
what do you mean by catalyst here? the target array Catalyst? still not clear, could you explain a little more?

[HannesK]

Hello Hirosh,
can you please describe what you try to do and what your actual problem is? This thread is two years old and I would like to be sure, what the real issue is.

Catalyst is the protocol that HPE is using for their StoreOnce products. If two StoreOnce systems should send data to each other via catalyst copy, then the protocol must match.

Example 1: source StoreOnce has only IP and target StoreOnce has only Fibre Channel... this will fail. The connection between the two StoreOnce boxes must support the same protocol. The is no FC to IP magic.
Example 2: backup job to StoreOnce 1 with fibre channel. IP connection exists between StoreOnce 1 and StoreOnce 2. "Catalyst copy" will use the IP connection between StoreOnce 1 and StoreOnce 2. That's all done automatically by Catalyst Copy on the StoreOnce side.

Best regards,
Hannes

[Hirosh]

Hi Hannes,

I have two sites which use the StoreOnce as their disk-based backup in the Veeam. the StoreOnce are connected to Veeam Servers over FC on each site and also as a target repository on other Sites Veeam Servers. we were planning to replicate the storeonce with catalyst copy to each other over 10G Ethernet. The issue is as a pre-requirement it is stated "This direct connection must be of the same type as the connection that you select when adding the target HPE StoreOnce".

Best regards,
Hirosh.

[foggy]

You can add the devices from another site via IP to allow that.

Just to make, sure, are you using different Catalyst Stores as repositories created by different backup servers?

[Hirosh]

Hi foggy,

each storeonce have 2 sets of catalyst stores, one for the local Veeam backup server backups repository and another as the destination backup repository for the replication from the source. am i being clear?

[foggy]

Yes, this is important as sharing a single Catalyst Store between different backup servers is somewhat of a challenge.