Start a Conversation

Unsolved

This post is more than 5 years old

874

February 14th, 2008 14:00

Stub Files Disappear - Moved Files 0 KB in Size

We've implemented DX on our main home server a few weeks ago. We've been conservative in our approach and currently are only scanning and moving on Sundays. Our move rules only move files whose last written to date is > 4 years. This has been successful thus far. We've freed up close to 100GB of storage and when we get down to adjusting the rules to moving files with a last written to date of >14 months we expect a big payoff.

We are having a Sev 1 issue (a case is currently open w/ EMC) our files that have been scanned and moved (those that up till now were stub files) can't be opened. Symptoms are the following:

-file appears to exist
-size of moved files are 0 KB
-opening these files yields a blank file (bet it a .doc, .xls, .mdb, etc)
-the target symbol in the bottom left hand corner no longer exists

I've done the following:

-stopped/started the service
-confirmed the data exists on the 2nd tier storage (using NAS)

What can be going on here? We have NOT changed anything in our environment. The last entry in the event viewer was on Monday evening stating that a file was fetched from 2nd tier storage. Nothing since then other than when I restarted the service earlier.

What could possibly be going on here?

397 Posts

February 16th, 2008 12:00

It appears as though the DXspy services are not loading. These services perform the file size fixup (so the stub appears to have size in explorer). They also tell explorer to display the shortcut icon. Enable trace logging in DX and restart the service. Post the log results here. It should indicate the source of the problem. My best guess is the volume id of the extended drive has changed. Good luck.

129 Posts

February 22nd, 2008 05:00

As I replied to you in "Server with newly installed DiskXtender 6.2 for Windows crashes", we experienced the same in Win2003 MSCS cluster environment.

After a while, we got 0 KB files, and when we tried to restart the service, it either crashed the server or did only a failover to other node.
It seems there is a memory leak in DX and the fix will be available in DXs SP4. Saddly not released yet.

45 Posts

February 29th, 2008 08:00

First thing to look at is your backup solution. Are you using VSP? If you take a snapshot with the Veritas Snapshot provider it creates a "virtual drive" with the same volume serial number and DX doesn't yet know how to determine which drive is the right one.

The only way around this issue until Q4 is to disable your VSP driver and use VSS instead. I ran into this a couple of months ago and disabling VSP did the trick.

4 Posts

April 8th, 2008 14:00

Ryan you were right on the money. I talked to EMC and Veritas and both confirmed the same thing. I changed from VSP to VSS for backups and I've never seen this issue ever again.
No Events found!

Top