7.1 Windows Clients and "No error" message 1

[kcongdon]

Server: Solaris 8, Networker 7.1.0
Clients: Windows 2000, Upgraded from 6.1.3 to 7.1.0
Compressed on the client side

We have 10 Windows 2000 servers that used to pick up all their incremental and full savesets just fine using the 6.1.3 client. The group notifications declare the backups to be successful, but make the following notations on some savesets:

U:\Server Documentation\Backup of Docs\: No error

[The client's savesets are C:\ and U:\]

Some of our full backups for these "No error" clients have dropped from 500 GB down to 65 GB. We know there's data missing because our Windows administrators can't restore files that should have been backed up. The data didn't start disappearing until the clients upgraded to 7.1.

We had a client try a full backup from the client side and all the files got backed up.

Any ideas on what might be causing the problem?

Thank you.

K. Congdon

 

[605]

First of all, check what exactly will not get backed up. Is it a certain partition/
directory or is it random?

If it is always the same data, i think it can easily be repaired as there are
only two pssibilities:
- either save sets or
- directives
Check both, server side directives and client side directives!!!!!!!!!!!

It is rather impossible to tell you why this happed. However you should be able to repair it.
On the other hand there is a new feature with NW 7 called Resource Update Logging.
Once enabled it will tell you who has changed resources and what exactly has been changed.
You might want to turn it to prevent potential problems.

 

[kcongdon]

605, thank you for your suggestions.

On one of our file and print servers, the full backup picked up only one subsubdirectory. It skipped everything else in D:\.

Have there been changes to security or how Networker reads savesets in 7.1 that might force us to more explicitly define the savesets?

 

[605]

No nothing has changed to my knowledge. And if so, i am sure there would be many
more statements like yours.

Look for local directory files (nsr.dir) in the root directory. If somebody tested this
feature and forgot to delete it again, the described may happen. If not, it seems
that there is a bug.

 

[Greenli]

I had this same problem and what I discovered is that the Local Administrator for the server did not have rights to the directories that weren't getting backed up. These rights got changed when one of the administrators setup a share and removed the Local Administrator from the file / directory permissions. With the 6.1.3 client we didn't have this problem so the way Legato deals with rights has changed from 6.x to 7.x. I simply added the permissions and the backups happened correctly again. Legato Tech support is aware of this problem and even have a "fix" for it. You can get the "fix", which is just a liblocal.dll file (Reference LGTpa49089), change the permissions, or put the client back to 6.1.3.

 

[keefb]

Hi All,

I have upgraded to 7.1.2 after having the same problem as kcongdon and am still having the same problem, this issue may NOT be resolved in 7.1.2, but should be fixed in 7.1.1.

I'll try and update the thread when I hear back from Legato support people.

Thanks

Regards
KeefB

 

[kcongdon]

KeefB,

You may also need to fix the folder permissions. Fixing the folder permissions made the error go away at first. The error kept recurring after the folder permission fix, so then we applied the patch.

Our Legato tech support vendor gave us a bug ID of LGTpa60860. The permanent fix was to upgrade to 7.1.1.

It's a little disturbing that a 7.1.1 fix didn't make it into 7.1.2.

Is the directory generating the errors empty? We're seeing errors on empty directories specified as separate savesets. The folder permission and patch combination has not fixed the empty directory problem.

Thank you.

Kate

 

[mwellons1]

It is the folder permissions and it has been fixed if you upgrade to 7.1.3

Legato says they put out a notice to all sites that this was an issue even though I never saw it come through and we just ran across this issue.