Rman Plugin V10

[isaez]

Hi,

I did migrate the Rman plugin from v9.5 to v10 and have a problem that can't be solved: I did configure rman to use a .bak extension for the rman backups and in v10 the .vab extension is compulsory. So if I try to delete a obsolete rman backup file which has the .bak extention the v10 plugin can't find it! V10 is appending .vab so it's looking for .bak.vab files.

I did open a case and this was the final finding:

I also rechecked the User Guide for Plugins v9.5, and found there this:
You must use the .vab postfix on any backup files created on the Veeam backup repository. Otherwise, Veeam tools will not be able to identify the backup files and perform backup import and other functions. It is recommended to implement the described above naming rules and use the repository ID postfix.

Plugin v9.5 does not enforce .vab extention, yes. But still we can not guarantee that everything will be working, that's why it is outlined in the User Guide. Can you please tell me, did you ever try to perform a test restore of from the .bak files, and if so, was it successful? And how often is the situation when you need to restore data from a months old DB? Also, can you please tell me. did you already updated plugins on production database hosts?
If you are sure that it is working, I would suggest moving those backups somwhere (on another disk, or on another server) for a year, remove the backups from the Veeam B&R configuration, reconfigure the RMAN to name them properly, and start a new backup chain. In case that you would need old backups - deploy Veeam B&R v9.5, install a plugin v9.5, import those backups and perform a restore. But I still need to point out, that in case it would not work, it would not be supported, since it falls under the limitation that is outlined in User Guide. So my point here is that the best course of action is to switch to the supported scenario as fast as possible, minimizing the possible impact.


So if you are going to upgrade to v10 be careful because if you don't use the .vab extension you could fase the same issue.
Upgrading to V10 was more or less mandatory for us because in v9x removing obsolete backups could table some 10 minutes per backupset.

regards,
Ivan

[PetrM]

Hi Ivan,

Yes, this is expected behavior.

So my point here is that the best course of action is to switch to the supported scenario as fast as possible, minimizing the possible impact.

This is good idea and I always suggest to use supported configuration only.

You may try the following steps to workaround the "delete obsolete" issue:

1. You need to update veeam_config.xml and set the following parameter:

Code: Select all

<PluginParameters shouldAddExtensionToFilename="false" />

The path to the configuration file is: "/opt/veeam/VeeamPluginforOracleRMAN/veeam_config.xml" in case of Linux OS or "C:\Program Files\Veeam\VeeamPluginforOracleRMAN\veeam_config.xml" if you have Windows based server.

2. Then do "delete obsolete". The lookup will be performed based on file name in RMAN catalog without additional .vab postfix

3. Once cleanup is done, please remove the parameter from veeam_config.xml or set true as parameter value:

Code: Select all

<PluginParameters shouldAddExtensionToFilename="true" />

This step is important because different functions which are related to the plug-in look for backup files using .vab extension.

It should do the trick but I recommend to ask our support team to help you with this procedure.
Please do not forget to paste case id for our reference.

Thanks!

[isaez]

Petr,

Great! I'll try it tomorrow first thing. Thanks a lot.

regards,

Ivan

[PetrM]

You are welcome!

Anyway, I suggest to ask our support engineers to help you with the steps above.

Thanks!

[isaez]

Petr,

I did implement the shouldAddExtensionToFilename="false" option and it works fine.
The version 9 plugin was also able to remove obsolete backups but it took +/- 5 minutes per file! So we upgraded to V10 and it works as it should be.
Now I have some orphan files (not longer in the rman repository) on the Veeam repository and after removing these files I'll set shouldAddExtensionToFilename to "true".

regards,

Ivan

[PetrM]

Glad to hear that it worked!

Thanks!

[isaez]

Petr,

I did implement "OracleRMANConfigTool --set-force-delete" to get rid of orphan files but it's not clear to when Veeam does remove the files. Is that once a day/week/or.. ?
And I would appreciate it if you could point me to the documentation of parameter like shouldAddExtensionToFilename. If I search the Veeamm website I can't find nothing.
Some technical documentation would be nice.

regards,

Ivan

[PetrM]

Hi Ivan,

The garbage collection task is triggered automatically after every successful backup and removes files which are older than the specified period: the number of days entered during "--set-force-delete" configuration.

There is no description of "shouldAddExtensionToFilename" parameter in the documentation because it's required for the specific cases only, however I think now that it's a good idea to add a description in our documentation. Thank you so much for the feedback!

Thanks!

[isaez]

Petr,

Thank you again! I' like very much the fast and good support you guys give. I did check the orphan files this morning and they were gone. Great!
Regarding the documention: I suggest you also add the information you gave me about how the garbage collection (--set-force-delete) works to the documentation. What can be found now is very minimal in my opinion.

And I had a problem with the rman veaam backup of a database. And I'm just sharing this so others can benefit from it> I go this error message:

Code: Select all

ORA-19506: failed to create sequential file, name="DATA_XXXXCDB_1540626645_20200313_p6ur11al_1_1.vab", parms=""
ORA-27028: skgfqcre: sbtbackup returned error
ORA-19511: non RMAN, but media manager or vendor specific failure, error text:
   Cannot convert string value to signed integer. Given string: [/u01/app/12.2/bin/oracle].
--tr:Failed to invoke command
--tr:Failed to invoke command: SbtBackup

I did search for the error on Veeam and Google but was not able to find anything. After an hour of checking and trying I did remove the file in /tmp/veeam_plugin_locks but the issue remaind. And the I saw ther were some old veeam process running on the server:

Code: Select all

serve:/tmp/veeam_plugin_locks $ ps -ef|grep 6791
oracle   12583 12491  0 Feb28 ?        02:51:24 RMANPluginManager 6791 /opt/veeam/VeeamPluginforOracleRMAN/RMANPluginManager
oracle   23855  6674  0 08:30 pts/2    00:00:00 grep --color=auto 6791
server:/tmp/veeam_plugin_locks $ kill -9 12583
server:/tmp/veeam_plugin_locks $ ps -ef|grep 12491
oracle   12491 12372  0 Feb28 ?        00:00:01 oraclexxxxcdb (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
oracle   12492 12491  0 Feb28 ?        00:00:00 [RMANPluginManag] <defunct>
oracle   12583 12491  0 Feb28 ?        02:51:24 [RMANPluginManag] <defunct>
server:/tmp/veeam_plugin_locks $ kill 12491

The process running where from the rman plugin version 9. And yesterday I did update to v10. But I did not notice the defunct process. So after killing the process the backup did run fine.

regards,

Ivan

[veremin]

Just so you know you can always send the feedback directly to technical writers by submitting your comments via "Send feedback" form available at the bottom of every User Guide page.

They are pretty quick in fixing issues.

Thanks!