This backup set may not contain any data

[Popolou]

Hi all,

What are the common troubleshooting steps in diagnosing the problem above? It seems that after a rather smooth few weeks since i carried out a hardware transition, only in the last few days has one logical drive on a server decided not to backup successfully.

BE is returning the rather generic error of This backup set may not contain any data and my KB research only turns up SQL and Exchange issues. Nothing relating to simple files and folders.

My efforts in resolving this includes disabling the Active File Exclusion filter via the registry (without luck) and running a B2D job (which works fine). Funnily enough it happens to be the Program Files directory containing SQL and Exchange server, yet its just the files as opposed to any databases.

With the B2D job set to full backup, this passes without a failure. I haven't tried an incremental on the same B2D however i've a feeling it will also succeed. It seems that the problem is restricted to backing up to tape.

Recounting any changes made to the system, i simply asjusted the block writing/reading mode of the drive to improve efficiency. It may be bad timing but i'm not convinced the two are connected.

Any pointers?

Cheers

Pop

[Popolou]

Bump

Seems to be a very limited resource of information about this error.

Regards

Pop

[Popolou]

Another folder off a different server has just begun to report the same problem now for no obvious reason - is there no indication at all why this happens?

The two folders are completely unconnected and contain vastly different data, yet why it is that they eventually error with "This backup set may not contain any data" is leaving me confused.

Regards

Pop

[RoninV]

This may be a rhetorical question, and you may have inidcated it in your post, but are these "no data errors" appearing on differentials backups?

[Popolou]

Hi Ronin,

Actually it's an incremental. My tests have shown that when performing full backups, everything passes as expected. However when an incremental is scheduled, these unique and mutually exclusive folders are being singled out. It seems the folder hierarchy is saved yet the data within is not.

To expand on the original problem, we included a new server in the backup job each night. One drive (the whole E: drive no less) hits the same problem for no apparent reason.

I haven't been able to put any time on this tbh, but i assume that i would need to troubleshoot if the archive bit is being reset or not.

Regards

Pop

[RoninV]

Pop,

The reason I asked, I'm getting the same "error" in my differential backups, though my full backups are fine. At first glance, when it comes to differentials, it would be easy to think, no data because there were no changes since the last full backup. Of course, I test that theory, by dropping a few data files on the "error" prone areas (hard drive of two servers), but that didn't solve anything.

I've read other threads which refer to a registry edit (Active File Exclusion), but AFE is described as an Exchange excluder. I need to backup Exchange as well.

[Ken_Putnam]

Actually AFE excludes known database types from flat file backups (assuming that you are doing an Agent backup of the database), and doesn't list a skipped in the job log

[RoninV]

So you're saying that AFS may solve the "backup set may not..." issue, if Exchange is on a separate server from the BEWS installation?

[Popolou]

My take on AFE is that it is mutually exclusive from the (backup set may not contain any data) error. My own tests as i described before, seemed to make the case in hand clear too. After all, it would be misreporting if ASE (being enabled as default) started to report that backup sets failed due to no data. Besides, i'm not convinced its an ASE issue as the environment remained unchanged form before and after the issue began.

I did a little investigating after my last post, but could only afford an afternoon. The first thing i did was a simple OS search to confirm when and what files were changed on the reported volume. Incrementing the search criteria by a single day each time, i managed to tally those dates when filesystem changes occured to what was backed up that specific night. Despite only a few KB changes in files, BE backed them up and did not report the error. However, it seems that there are long breaks when something on that volume does change and to when BE picks up on the changes which led to our initial concern.

This, at least for my benefit, confirms that the errors are informative only and that despite the file structure being comitted to tape (coming in at a mere 52Kb, say) it is not enough for BE to accept data was backed up. It would consider at least a single file backed up in order to not trigger this error.

The only other remaining "no data" errors relate to a \\BACKUP SERVER\Cont?Prot?Srv?Config fault which i assume relates to a snapshot problem in CPS and another volume (as above) that contains data that is not part of any backup selection, however.

It would be handy mind you, to be able to disable the error on certain drives/volumes where you know that data is not selected as part of any back up. BE, reporting the error when in actual fact it was configured as such, had our knickers in a twist until we clicked onto what was really happening.

Ronin, after you perform a full backup, carry out an incremental on the same selection once you either changed or modified (or both) a directory to test. Bear in mind however that AFE excludes other files such as SQL .mdb's too. For reference, our servers are performing a flat file backup of the exchange server directory successfuly, while the essential logs and databases are on a seperate volume and backed up via the usual agents. What's left over in the program folder doesn't seem to change much and hence the occasional problem.

Panic over i guess.

Regards

Pop