MaxDB Plugin

[ANDREAS_JANKOWIAK]

Hi all,

I've questions regarding the MaxDB plugin. Hope someone can help ;o)

1. MaxDB Plugin is no longer a community plugin, right? It's officially delivered with SC 3.6 ?

2. Which MaxDB version is supported?  Is it really just version 7.8 like stated in the SC 3.6 installation guide?   What about version 7.7 ?

3. Is there a special filesystem  / Volume layout needed? 

Thanks a lot for your responses.

Best Regards,

Andreas

[ktenzer]

Hi Andreas,

1. MaxDB is still available as community plugin. All plugins are available through community just if you use them you dont get NGS support. Only plugins that ship and are included with SC are supported by NGS. As of SC 3.5 MaxDB plugin is supported and ships with SC.

Yes it is officially delivered with SC 3.6

2. Please refer to the IMT

http://support.netapp.com/matrix

3. There is no special file system layout however it is strongly recommended to separate logs and db files on separate netapp volumes. In addition NFS makes things quite a bit simpler from a recovery / restore perspective.

Hope this helps

Keith

[ANDREAS_JANKOWIAK]

Hi Keith,

thanks a lot for your response. Unfortunately I have no access to the comp. Matrix ;o(    My user does not have the permission.

Do you have a document which describes a  restore of a snapped Maxdb?

Is there a description about the functionalety of the plugin?

Are these steps executed when snapcreator triggers the snap via the plugin ?

1. Write Maxdb Savepoint

2. Suspend_logwriter

3. NetApp Snapshot

4. Resume_logwriter

Thanks a lot for your help!

Best Regards,

Andreas

[ktenzer]

There is a TR on MaxDB and NetApp but it is pretty high level

http://www.netapp.com/us/library/technical-reports/tr-3689.html

Solutions builder which is a tool partners and netapp have access to has the detailed documentation and procedures. Since not sure if you have access I have attached DOC for MaxDB and SC.

Snap Creator will restore the data but not perform recovery. You need to preform recovery of MaxDB manually after doing restore in SC. If you want you can automate this through SC using RESTORE_POST_CMDs. Again recovery procedure is attached in document form solutions builder.

Keith

[ANDREAS_JANKOWIAK]

Hi Keith,

thanks a lot for this document!

Can you confirm that if the maxdb plugin triggers a snapshot, a maxdb savepoint is triggered?

And it would be very helpfull for me to know if Maxdb 7.7 is supported by Netapp or not. We can't upgrade to version 7.8....

Thanks again.

BR,

Andreas

[ktenzer]

Yes SC will take a consistent maxdb backup which depending on version involves a savepoint.

Only MaxDB 7.8 is supported in IMT

7.7 shouldnt be a problem and certainly you can contact your NetApp account manager and a PVR could be filed to get support.

I will ask QA why 7.7 isnt supported, probably was just testing resources.

If you have a problem you can of course contact Netapp support, they will always try and help.

Regards,

Keith

[ANDREAS_JANKOWIAK]

Hi Keith,

thanks a lot, it would be wonderful if you could ask QA about the version.

BR,

Andreas

[ANDREAS_JANKOWIAK]

Hi Keith,

I've tested now a little bit with the restore of the database.  Unfortunately I've some problems... maybe you have an idea.

Case 1:

Restore Data Volume --> db_restartinfo -->  Needed Pages are still in the LOG Area --> db_restart ---> DB online with latest state

Case 2:

Restore Data Volume and LOG Volume --> Needed Pages are not completly in the LOG area --> so I have to restore archive logs

--> recover_start <medium_name> log 005  --> after that the database goes online! But we don't have the latest state. 

Is there a way to prevent the online state and continue with next logs with recover_replace log 6  ?

I wonder why the database goes online... if I restore the log area from the backup, it can not contain the latest log information. So if the db goes online the page is not the one which is contained

in the latest log file.

Do you have any idea?

When you do a traditional maxdb restore via Backup Images, you would normally reinitalize the database and log area.... but when I restore the data and log vol from the snapshot it's the same I guess...

I'm a bit confused about this fact at the moment....

Thanks!

BR,

Andreas

Nachricht wurde geändert durch: Andreas Jankowiak Seems I have to clear the log area first with db_execute clear log. After that I can recover sucessully. Bad thing is, that the history log gets status histlost.

[ktenzer]

Unfortunately this is getting beyond my maxdb skills.

There is option I think in config under maxdb section to not do a savepoint, you can try that

Keith

[Norbert_Sommer]

Hi Andreas,

here's my view on some of your questions:

> Suspend_Logwriter

SnapCreator for MaxDB has a Y/N switch for that, you can choose.

Based on my research until now this would not be necessary for Netapp/NFS and if all data files are on the same volume.
For NFS with data files spread over multiple volumes there would be volume-groups.
I think the suspending of the MaxDB logwriter would be used in cases like a SAN where the storage controller has no visibility of
the file system which in that case is owned by the host OS.

> Automatic Savepoint
Not sure, but I think there is a feature in MaxDB to trigger an automatic Savepoint in less busy databases every <x> minutes.
You could always use a SnapCreator Pre-command for that or just test it:
there must be a DB table like V$Savepoints which stores this information.

> MaxDB Versions

I'm not involved is the development of the plug-in but I would expect it to work with all current versions like 7.6, 7.7 and 7.8.

> Case 2 - Point in time recovery

You executed a "recover log #" so I think it will at the end of that log, start the DB and this is what you requested.

If you need a point in time recovery which does not go to the very latest Savepoint (that's case 1) the syntax

would be something like "db_restart -until <date> <time>".

Best Regards, Norbert Sommer