Quantcast
Channel: All Data Protector Practitioners Forum posts
Viewing all 10494 articles
Browse latest View live
↧

Re: HP Data Protector 9.0 Backup to HP StoreOnce NAS Share

November 14, 2016, 8:41 am
$
0
0

On the 4210, is the NAS share permissions "CIFS authentication is managed through AD" ? If so, check permissions via running MMC, then add snapin, click "shared folders", ADD, then next window click "another computer" and enter D2D hostname/IP. Then click ok. Shared folders for the D2D should appear under "Console root" in the left hand pane. Click shares and sharename to get properties. Check share permissions there....

some screenshots would be good to assist.

also what DP inet is running as plus what ownerships you have on the backup spec....

Can you browse the NAS source shares from the DP GUI source tab of the backup spec?

↧
Search

Re: HP Data Protector 9.0 Backup to HP StoreOnce NAS Share

November 14, 2016, 8:59 am
$
0
0

Ok so when I open the MMC, browse by IP to the share, it connects but says access denied which is strange since I'm a domain admin on this domain. From the DP side, I can see the device and the slots, but it says they are empty.

↧
↧

Re: Exchange 2016 Support

November 14, 2016, 11:58 pm
$
0
0

There is limitation for DP 9.06 support for exchange 2016:

 

- Microsoft Exchange Server DAG configured without a Cluster Administrator Access Point is not supported.

With DP 9.07 there is some enhancement but still not up to the point:

- Microsoft Exchange Server DAG configured without a Cluster Administrator Access Point is supported with Round Robin DNS    mapping of DAG name to all the node IPs.

I'm still searching for whitpaper or document which specifically explains how the setup will be done.

↧

Re: HP Data Protector 9.0 Backup to HP StoreOnce NAS Share

November 15, 2016, 2:01 am
$
0
0

ok, check the NAS configuration on the storeonce under the "CIFS Server" tab and the members. Maybe add a group there? Obviously a permissions issue if you cant see the shares via MMC....

sorry its not the devices, go into the backup spec in the source tab. Can you drill down the network shares?

↧

Re: Should we have a DP startup command in our ServiceGuard script?

November 15, 2016, 3:58 am
$
0
0

Your mixing up IDB corruption and node failovers

If your IDB should get corrupted on node A, failing over to node B is not going to make things worse.

Anyway your DP-package, served by node A or node B, == the Cell Manager == catalog DB owner

(omnidbutil -show_cell_name)

What I should at to my previous entry:

 /DP9/etc_opt_omni_server/sg/csfmodular.ksh monitor --> was a call to a slightly improved script for DP 9.02

My experiences with DP 9.05 showed me the default script in ..../sg directoy was sufficient.

(I guess csfmodular.ksh which I developped with Zoltan Tuba was fitted in the newer releases  :-)  )

--> simply use the standard cell server script but be aware the implementation guidelines for ServiceGuard  in the manaual are bogus, but that you must have discovered by now.

Luc

↧
↧

Succesfull session in database, unable to restore from that point in restore section

November 15, 2016, 4:45 am
$
0
0

Hi guys,

 

I am trying to restore certain folders from a server. In the internal database I can see the succesfull sessions for that specific drive of the server. In this case 2 February 2016.

Now when I go to the "Restore" tab to restore certain folders from that drive, the furthest point I can go back to is 28/06. I can't see any reason why the other restore points are not there.

Has anyone else encountered these problems?

↧

Migrating Windows Cell Manager IDB

November 15, 2016, 5:45 am
$
0
0

Hello!

I did a lot of tests to migrate a Windows Cell Manager IDB to a different Server, as described in the Installation Guide DP 9.08 Page 285 ff.
OLD_SERVER was a Windows Server 2008 x64 with DP 9.08 , the NEW_SERVER is a Windows Server 2012 R2, both german OS version.
The NEW_SERVER name is different from the old one, the IP address is the same.
The last successful FULL backup of the IDB is stored on a LTO3 tape.
On the NEW_SERVER I created an LTO3 backup device identical to the old one and successfully imported the media.
After importing the IDB ( page 289 ) there are lots of problems:
- I must definitely replace all keystorePassword and truststorePassword und generate a new certificate ( page 291 ) to connect to the CM
- Clients are missing
- only filesystemsbackup specifications are listed, Exchange 2010 is missing , schedules are missing
- only NEW_SERVER is listed as a client, but LTO3 device ist owned by the OLD_SERVER

It is to mention that the documentation is written very careless ( like /etc/opt/ ... for a Windows Server ) and one point is missing:

the clean-up after an successfully migration , because the db80_restore\idb path keeps the active path !!

Any ideas how to solve my trouble?

Best regards, Dirk

↧

Re: Should we have a DP startup command in our ServiceGuard script?

November 15, 2016, 6:05 am
$
0
0

Thanks Luc!  I was mixing IDB corruption with the failover process.  You addressed my concern about potentially causing further damage in your last post.  Thanks for clearing this up for me.

Kind Regards,

Brian

↧
Search

Re: Succesfull session in database, unable to restore from that point in restore section

November 15, 2016, 6:44 am
$
0
0

Hi,

given 28/06 is further away than 3 months, I assume you already changed the Search Interval box to None from the default 3 months? You should see your objects then, provided

  • They were actually part of the backup at this time
  • The session is still protected or at least not overwritten (should be, given you still see the session log)
  • The session is still under catalog protection

With the search interval set to None, in the displayed tree find a file or directory that should have been there from the beginning (like lost+found on Unix or the $Recycle.Bin on Windows), right click it and choose options. In the dialog that pops up, check the Version tab. Is there a version of that file/dir in the Backup version dropdown (or the more detailled version selection dialog hidden behind the "..." button)?

HTH,
Andre.

↧
↧

Media object copy from command line; how?

November 15, 2016, 7:00 am
$
0
0

I wrote a script that selects from all protected media those who have less than 50% of the used capacity protected (Data Protector cannor do that!). Now I'd like to create a media object copy job by script that copies all the protected objects form the media found in the previous step to another media (to free the source media completely).

It seems to me I can only run such a copy interactively from the GUI (or the manual pages are way too incomplete).

Can anybody tell me how to start such a command from command line? I'm using Linux for the Cell Manager...

↧

Re: Succesfull session in database, unable to restore from that point in restore section

November 15, 2016, 7:09 am
$
0
0

Hi André, 

First off, thanks for the quick response. I had already set the intervals right.

I can also see the objects. But when I open the versions tab, the oldest version is 28/06. Even though in the IDB it says that there are succesfull backups from:

- 2016/02

- 2016/01

- etc.. even older.

 

I can also see to which tapes the back up has been written to, and these tapes are in the connected library on this moment. Is there any other way I can try to directly restore from the tapes themselves?

 

Kind regards

↧

Re: Got unexpected close from RMA

November 15, 2016, 7:11 am
$
0
0
GrassHopper wrote:
Nov 9 10:46:14 ret-rh1p kernel: rma[55085]: segfault at 7fc84ef7c000 ip 00007fc84f4343dc sp 00007ffd98f9de30 error 4 in libserializer_64bit.so[7fc84f408000+52000]

Having the RMA segfault in the serializer lib (either during copies or during restores) for StoreOnceSoftware objects that otherwise succeed a Verify is exactly what I had here, after upgrading to 9.06, but on Windows. I've been working with HPE on a case for that and finally was told the bug in question should be fixed in 9.08. I just upgraded and indeed, the very same objects that made the RMA crash before are now copied without any fuzz. So if you are on 9.06 or 9.06_108 (dunno about 9.07), upping to 9.08 might be a good way to get rid of that crash (provided we are talking about the same final cause).

HTH,
Andre.

↧

Re: Data Protector Cell Manager

November 15, 2016, 7:27 am
$
0
0

Hi,

moving the CM to another machine is quite an involved operation (where we define move as moving the installation to a box with another IP address, hostname and maybe file system structure), while moving away the MA is trivial. As such, you might rethink what you are planning. If the intention is e.g. to P2V the original CM+MA box so the CM becomes a virtual machine, while the physical box is reduced to a media server for the tape libraries (or something like that), you may still consider to P2V the entire machine keeping hostname, IP address and file system structure, then rebuild the physical box with a blank OS and finally move the MA away from the VM to the "new" physical media server.

HTH,
Andre.

↧
↧

Re: Number of Detail Catalog Directories limit reached DP5.5

November 15, 2016, 8:28 am
$
0
0
higo wrote:
you are right, we never purge the DB and we are afraid to do so at the risk of losing data.

In your specific case, purging alone will not help with anything and rebuilding the IDB after purge will only reduce your fnames.dat* footprint, but you are hitting the DCBF directory limits. These have been bumped up in later versions and the entire IDB became way more scalable in 8.0+ but given you are still on 5.5, upgrading through that chasm of time and code churn isn't going to be an easy feat.

Is there any way you can get rid of some protected catalogs? That is what blows your DCBFs now. If you really have a large pool of media kept intentionally at Permanent (or otherwise very long) protections, consider reducing catalog protection for the oldest stuff somewhat - it can be reimported if need be, but it doesn't have to fill your disks in all eternity.

HTH,
Andre.

↧

Re: DP 9 Client installation = Could not use secure shell protocol

November 15, 2016, 8:56 am
$
0
0

The FQDN is what did it for me !!!

 

Thanks

↧
Search

Re: Media object copy from command line; how?

November 15, 2016, 12:07 pm
$
0
0

Hi Ulrich,

please refer to the CLI reference guide (CLIReference.pdf in the Docs folder). It contains details on the omniobjcopy command that can be used for Object Copy in a script.

Please use the Accept Solution button next to my post and assign a KUDO (thumbs up icon) if this works for you.

Regards,
Sebastian Koehler

↧

Re: Should we have a DP startup command in our ServiceGuard script?

November 15, 2016, 12:43 pm
$
0
0

Hi Brian,

I see  was able to help you on this topic.

Based on my experience any running services should be integrated into the package unless the team working on the cluster really knows what is going on and what to do when. If MC/SG or SGLX finds an issue it can't resolve it will simply not being bring up the package (on the other nodes) and scream for help. Starting and monitoring the Data Protector services with Serviceguard offers the best protection against unplanned service outages. It should be documented and tested how the package is put in and out of maintenance mode. Starting services manually always has the potential to bring the package in a state where it requires manual cleanup before it can be reactivated.

To be honest, I'm still not convinced that the supplied scripts will do a good job for modular style packages. This was the reason for putting together my own scripts to start, stop and monitor the Data Protector services in the cluster years ago and they still work very well.

Please use the Accept Solution button next to my post and assign a KUDO (thumbs up icon) if this works for you.

Regards,
Sebastian Koehler

↧
↧

Re: Media object copy from command line; how?

November 15, 2016, 11:08 pm
$
0
0

When I wrote "or the manual pages are way too incomplete" please don't just tell me to read the manual page when you don't know the solution!

↧

Re: Media object copy from command line; how?

November 15, 2016, 11:27 pm
$
0
0

The manual is not limited, it is the omniobjcopy command itself. Most of the Object Copy operations that can be executed from GUI can't be done from command line. If you're not able to make it work with the available command switches you need to run it from GUI. It is simple as this.

FYI, BRICK has a module called space claim. Not sure if it already generates the necessary omniobjcopy commands, but it might be worth looking into it.

Please use the Accept Solution button next to my post and assign a KUDO (thumbs up icon) if this works for you.

Regards,
Sebastian Koehler

↧

Re: Media object copy from command line; how?

November 16, 2016, 12:41 am
$
0
0
Ulrich Windl wrote:

I wrote a script that selects from all protected media those who have less than 50% of the used capacity protected (Data Protector cannor do that!). Now I'd like to create a media object copy job by script that copies all the protected objects form the media found in the previous step to another media (to free the source media completely).

I reckon that's not explicitly provided by DP because it's a situation better to be avoided in the first place. It may build up occasionally by accident (wrong protection on some backups/copies, fixed later) or due to churn in long term protected media (objects were protected for years or even permanently, then some part of the objects became irrelevant and their protection was revoked or shortened). In such cases, the Interactive Copy GUI is usually enough to fix it. I would not go to the extreme of trying to automate this (except maybe when I inherit a heap of such media that grew out of bounds already, but not to be part of regular backup ops), instead to fix it at the source. Objects with different protections should funnel into different pools and if that means there are a lot of them, consider using a free pool to simplify the media lifecycle. Or unify the protections into less classes than you had before. Or both.

It seems to me I can only run such a copy interactively from the GUI (or the manual pages are way too incomplete).
Can anybody tell me how to start such a command from command line? I'm using Linux for the Cell Manager...

Ignoring the manpage for a second, I had a look into omniobjcopy -help and most of what I know from the GUI seems to be reflected there one way or another. So if you already have a list of objects you need to copy, the format

omniobjcopy [GeneralOptions] Object [Object] ... [Device] ...

looks quite right. You apparently have to know the individual object's type, name (a tuple specific to the type), session ID and ideally copy ID to make the selection unique, but that's not really different from the GUI methinks. What's missing?

HTH,
Andre.

↧
Viewing all 10494 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>