Import remote catalogue fails

[alfieboy]

Dear all,

I have started to investigate this problem, however there is a sense of haste and I would like others opinions in the meantime.

Background: Two Linux appliances, Netbackup 7.5.0.5, are attempting to "Initate" and the "Import" one remote catalogue from one remote Solaris 9, Netbackup 6.5.0.5 servers. The communication link between the backup servers is fine.

There are two separate "faults".

In the first case, the 7.5 Netbackup server starts the initated import and completes succesfully. However it does not list any catalogues for phase II, the import stage. The log file does report a series of INF and WRM messages in the log file stating existing image information has been found and this image cannot be imported, respectively.

In the second case, the second backup server completes phase 1 successfully and again confirms this in the activity monitor and the log file. However, there are no catalogue entries listed and when I attempt to search, using a start and end date, the system reports no images matching the selection criteria that were ready for phase II.

I have attached two log files, they are named accordingly.

I have just started to undertaken this work without any prior detailed historical knowledge of the backup policy at the time, therefore I am not familiar with start dates. Any quick thoughts would appreciated.

Best regards,

Alf

[SymTerry]

Hello

To make sure I understand, your tring to import a 6.5 catalog backupto two 7.5.0.5 appliances? Or is that a typo and your at 7.5.0.5 on all your systems.? When doing a catalog recovery, it is required that your on the same version. 

Also are you doing a catalog recovery?

[alfieboy]

Hello SymTerry,

Thanks for your commnet. Basically, we have two Netbackup 6.5 servers and we want to import the contents of one tape from each of the 6.5 servers to two 7.5 appliances. Each appliance server to contain the contents of one of the tapes from the 6.5 servers.

I hope this clarify the point.

I understand from the documentation, it is possible to impiort the contents media to another server. In our case, we only require the import of one tape per site from the old 6.5 servers to the replacement 7.5 appliances. The quantity of data is not great but we do need to retain this data set per machine.

Regards,

Alf.

[Mark_Solutions]

In your log file they all say this:

15:32:24 WRN - Cannot import client hostname, policy Policy_Names, and schedule Policy_Names_monthly_Full, copy 1 fragment 1 already exists. Skipping this image.
 

So everything on that tape is already known to NetBackup and so nothing new has been found on that tape to import

It has completed successfully but there is nothing for NetBackup to do.

[alfieboy]

Hello there Mark,

In my haste to request advice I have omitted a satisfactory description of the problem. I shall recap from the start.

1. Two geo-sites

2. One Netbackup 6.5 Solaris instance at each site

3. One Netbackup 7.5 appliance instance at each site.

4. Netbackup 6.5 still running in parallel with the 7.5 kit per site.

5. Each Netbackup 7.5 instance manually configured to remain consistent with Netbackup 6.5 regarding Policies, Volumes, Hosts etc..

6. Netbackup 6.5 using tape media

7. Netbackup 7.5 using disk but has a tape attached which is not used at present.

Objective:

The objective is to import a tape image from an individual media on the Netbackup 6.5 to the Netbackup 7.5 system per site. I have carried out phase 1, whereby the Netbackup 7.5 is configured to communicate with the Netbackup 6.5 in order to image the tape. On each ocassion the "Initialise Import" completes with a status of 0. There are no reported error message in the log file.

However, no images appear in the "Search results" window pane of the GUI. I use all standard search criteria apart from the "Start" and "End Date" fields, where I expand the search back a couple of years.

I have a feeling the fact that both sets of Netbackup instances (6.5 and 7.5) are virutally identical in terms of configuration, policies, hosts, volumes this is causing the problem. When the Netbackup 7.5 servers read their respective tape media, they find their own configuration is the same. If this and why it causes a problem I have yet to understand.

I continue to research this issue myself, but if you have any thoughts I would appreciate passing them on.

Best regards,

Alf.

[Mark_Solutions]

The results of your phase 1 that you showed us indicate that the images are already known to NetBackup.

I cannot see how it can be caused by identical setups as the images do actually have an image id as well as name and policy

Try just typing bpimport from a command line on the master and see what happens

[TomerG]

Also, if the master server on a site thinks that the image already exists (i.e. is already imported / in the configuration), this is easy to test: Try a restore for that client and see what files it thinks it can restore.

[alfieboy]

Hello Tomer,

Thanks for your comments. No, unfortunately. The Netbackup 7.5 appliance only contains data backed up using the Netbackup 7.5 appliance. There is no reference to the any earlier data which was backed up by the Netbackup server 6.5.

Regards,

Alf.

[Marianne]

Found existing image information for client hostname, policy Policy_Names, and schedule Policy_Names_monthly_Full, done on 03/02/2014 03:28:29.

The above means that a previous import attempt was probably successful.

Please open a command shell on the Appliance, 'elevate' to superuser level and issue this command:

bpimagelist -client <hostname> -d 03/02/2014 -L
Please show us output.

About the following: 

17:07:28 INF - Initiation of bptm process to phase 1 import media id Media_ID was successful.

It means that imported has started. Nothing else. Tape has not even been mounted yet.

Look at the output of the other import attempt - you can see the 'Initiation' message is followed by 'Waiting for mount....'

If you are doing imports from a remote Admin Console, take care NOT to close the console before the import is finished. It will kill the import.

Check bptm log on the appliance to see if anything is logged for the import attempt after 17:07:28.