Moving restore points from one repository to another? (StoreOnce Catalyst stores)

[evilaedmin]

Greetings,

I'm running into an issue with performance on an HPE StoreOnce: backups are fast, restores are reasonable, but StoreOnce Housekeeping is falling behind.

One of the recommendations given is to create more Catalyst Stores and spread backup jobs around.

Is there any atomic operation I can do to move a job and its related restore points from one location to another?

Thanks!

[jmmarton]

Eugene, about the only way today to move things is to put Catalyst shares in a SOBR and then use maintenance mode with evacuation but that will move everything off the old store. There isn't a way to manually move individual jobs, at least not with VBR. HPE has a tool though which may work.

https://h20392.www2.hpe.com/portal/swde ... ceCATTOOLS

Coming in v10 it will be possible to leverage "sealed mode" to at least prevent backup jobs from writing to an extent while still otherwise keeping it online. That will be an option as you could put the new shares into the SOBR and just seal the old extents. Just keep in mind that depending on how many shares HPE is recommending, you will have to be aware of SOBR licensing limits: Enterprise allows for two SOBRs with up to three extents each while Enterprise Plus allows for unlimited SOBRs with unlimited extents.

Joe

[evilaedmin]

Joe,

Thank you, I'll keep the catalyst copy utility in mind. Since our retention period is a pretty narrow rolling window, could I simply clone the subset of jobs I want to move with a different target repo location, disable the old jobs?

I don't mind trying to manage our retention manually on the old jobs.

[jmmarton]

Yup, cloning the jobs is always an option as well. The chains will start all over again meaning the next backup is an active full, but then again sealed mode in v10 would do that as well.

Joe

[evilaedmin]

It seems that stopping housekeeping and restarting it an hour later did something, as it's completed in under 24 hours with no change in activity.

[evilaedmin]

I don't mind trying to manage our retention manually on the old jobs.

Spoke too soon, this turns out to be not so easy.

I can identify the restore points easy with get-vbrrestore point. The findstorage() method gives me the FilePath in storeonce:// URL style. I THINK I can identify chains of backups with findpreviousobjectbylink() and findnextobjectbylink().

But is there a powershell method into Delete operation on storeonce://? This would be much easier than the Catalyst Copy utility, which I'm having some trouble with.

[evilaedmin]

Edit: Sorry, erroneous post.

[veremin]

The existing PowerShell cmdlets do not allows removing specific restore points - they will remove the whole chain instead. So, manual interaction via Catalyst utility seems to be the only valid approach. Thanks!

[evilaedmin]

Vladimir: Is it then appropriate to do "Forget" of missing restore points?

[veremin]

Correct, after you remove restore points manually and re-scan repository, you should use forget function to clear references from the product database. Thanks!

[evilaedmin]

The utility is very under-documented -- quick example: It supports the use of a --password-file argument but the format of the password file is not documented in the Doc!

Also not documented:

HPE community post

You will not be able to do that , 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.

Which is exactly my experience.

How may I submit an RFE to have Veeam make use of that field? Thank you!

[evilaedmin]

The existing PowerShell cmdlets do not allows removing specific restore points - they will remove the whole chain instead. So, manual interaction via Catalyst utility seems to be the only valid approach. Thanks!

Vladimir: Just to clarify: does "whole chain" mean all restore points or a single "unit of chain" (A level-Full and its dependent incremental) ? If so I can still use this method.

[veremin]

Haven't tested it in a while, but as far as I remember it removes literally all restore points, not specific subset. Thanks!

[ianbutton1]

Has anybody successfully used HPE StoreOnce Catalyst Copy Utility to move files? "Under-documented" is an understatement! Documentation for the StoreOnce Catalyst Credentials utility is in HPE SAP HANA Plugin User Guide https://psnow.ext.hpe.com/doc/a00022291en_us page 23. I've created the credentials file, but copy commands fail without explanation. Maybe logs can help . . .

[foggy]

Hi Ian, that's a sort of a question that should better be addressed to HPE than to Veeam. Take a look at the new backup copy capability in Veeam B&R v10 - it allows using Catalyst Copy to copy backups created by regular backup jobs between Catalyst Stores.

[ianbutton1]

Ah, here we are (my own notes) :-
StoreOnce Catalyst Copy Utility (Windows version)

1. Let’s assume that in the StoreOnce Management Console you can see a StoreOnce with IP address 190.100.100.10 with a Catalyst Store named Store1, and an admin-level user named User1. You want to copy catalyst files to StoreOnce 190.100.100.20, Catalyst Store Store2.
2. Download the utility from HPE, and run the exe to install it on the PC. It will create a folder structure, with two .exe’s in the bin folder, and a plugin.conf in the config folder. You will create a credentials file in the credentials folder called pass.txt.
3. Edit the plugin.conf file in BASIC CATALYST SETTINGS section, to set values for CATALYST_STORE_ADDRESS:190.100.100.10
CATALYST_STORE_NAME:Store1
CATALYST_CLIENT_NAME:User1
(Un-comment the lines and add the values)
Also add a line CATALYST_CLIENT_PASSWORD_FILE:pass.txt
In the CATALYST COPY DESTINATION section, set the following :-
CATALYST_COPY1_STORE_ADDRESS:190.100.100.20
CATALYST_COPY1_STORE_NAME:Store2
4. Create the encrypted credentials file called pass.txt by running the following :-
storeoncecatalystcredentials --add –u User1 –s 190.100.100.10 –o pass.txt
Enter the password for the user when asked, and again to confirm, then Y to save the file.
Repeat, to store the credentials for the second StoreOnce
storeoncecatalystcredentials --add –u User1 –s 190.100.100.20 –o pass.txt
I believe it has to be the same username, and obviously must exist in the Management Console of the second StoreOnce.
To check the entries, type storeoncecatalystcredentials --list –o pass.txt
5. Run Copy Utility commands, such as
storeoncecatalystcopy --copy -username User1 --preview
Any options not specified in the command will be picked up from the plugin.conf file - i.e. StoreOnce source & destination IPs, source & destination Catalyst Store names, password file location (to pull in passwords).

NB Be careful – always use preview option to check what will be copied/expired/deleted.

This is still work-in-progress - it lists some Catalyst files, but not Veeam files yet. (I saw some comment about a "completed" flag being necessary, but not implemented by Veeam.) But the idea is to pull Catalyst files into a firewalled StoreOnce protected against ransomware.

More anon, hopefully . . .

[foggy]

Does the task cover anything stored there or just the Veeam backups?

[ianbutton1]

Ideally I'd pull everything stored there - I can see our Data Protector backups (which have "completed" in the taglist). But Veeam's invisibility is a problem. In case we can't get hold of another StoreOnce for the "secret" side, I'm thinking of using other "pull" methods (e.g Robocopy) to get non-VM data across onto a secret NAS, and using Veeam Cloud solutions to protect VMs (with the benefits of Insider Protection and Immutable Storage).
SOCCU is supposed to be pretty much for L4 engineers only, but copying Catalyst items isn't dangerous to the source. If we can pull everything across to safety, we can work out how to recover chains & backups later - probably just copy the Catalyst items back to the primary StoreOnce on the main network, once ransomware is evicted. The primary VBR SQL database might be a bit out-of-date, but we back that up too.
A pity that Veeam File Copy and Import Backup can't see into Catalyst repos - but that's the oddball nature of Catalyst, I suppose. We can't go to V10 yet because we still have VMs on ESXi 5.1 which isn't supported beyond 9.5U4 (never mind that it's long out of VMware support!).