Windows Server Backup [2012 R2] - Went from 100+ copies available to 1 copy available overnight

Question

I'm using Windows Server Backup on a Windows Server 2012 R2 system that has been in production for half a year or so. The server has around 60GB used, and the backup drive is a 160GB external drive [yes, it's small, but will hopefully swap later]. The setup has been working great, and just yesterday when I opened the Windows Server Backup utility, I noticed that all backups have been successful and there were well over 100 backups available. The external drive was low on space, with probably 15GB free or so. But that's fine, since I understand that Windows Server Backup should delete the oldest backup available.

...OR...you know, it deletes everything. So I log in today, and what do you know? 60ish GB used on the external drive, so plenty of free space. But when I checked Windows Server Backup, I only have 1 copy available. Am I missing something? Is that how it's supposed to work when the drive gets full, that it completely wipes all previous backups and creates a new one? I was under the impression it would delete files from the oldest snapshot, or delete several old snapshots, and create a new backup once the drive had enough space freed up.

In my searching before this, I found that some others have had experiences, at least with Server 2008 R2, where Windows Server Backup would all the sudden only keep 1 backup copy, and consistently only ever update 1 copy and not maintain history, as if the retention flag had somehow been set to 1 without the administrator doing anything. So I guess I'll know tomorrow whether or not that's happening to me; whether I see "1 copy available" again, or "2 copies available".

Is this normal behavior? Any way to fix it? Thanks in advance!

:: EDIT ::

So...I checked this morning, and the server is backing up extremely slowly. It usually takes an hour and 15 minutes or so, starting at 2AM. It's 11:17 AM, and it's still backing up.

:: EDIT ::

The backup eventually finished. I still get the odd I/O error on the external drive and I'm planning on replacing the drive. However, I'm still curious of what's caused it to start only keeping a single backup. The only thing I noticed is that the Advanced Options in Windows Server Backup for that backup schedule now shows "VSS Full Backup", where as another server shows "VSS Copy Backup", and that server is maintaining history and each backup says it's a full backup.

So! I'm thinking I need to: •Cancel the backup if it's still running [taking abnormally long] •Delete the backup schedule and 'release' the external drive •Replace the external drive and create a new backup schedule, using the advanced option of VSS Copy Backup.

Can anyone confirm that this will most likely be the solution to get multiple full backup versions going again?

Solution

it took us a few weeks and lots of trial and error to figure this out.

Despite backing up every day successfully only the previous day was restorable on one of our servers (2012) and the windows server backup GUI showed only 1 copy on the ISCSI drive. Recreating the backup from scratch (and prior deleting all shadow copies) did not help.

It turned out the assigned space for the volume shadow service for the specific drive (drive properties) was set way too low. After increasing that amount the backup has created several copies and we could now restore more than just a single day.

Best regards,

complingua

Comments

[Michael Hampton]

Is one of your disks failing?

[Spence]

What do the Windows Server Backuo log files look like?

[link470]

@MichaelHampton internal RAID member drive, or the external drive? I really hope it's not an internal drive, the RAID manager says everything is healthy. There's a hot spare however that is showing "predicted failure", but that drive isn't active and I'm ruling it out and replacing it soon anyway. Are you thinking the external drive used for backups is failing? This was my initial thought, but I'm not sure why Windows Server Backup would completely wipe that drive when it was full.

[link470]

@EvanAnderson the backup log in Applications and Services>Microsoft>Windows>Backup show nothing out of the ordinary. A single 'Backup has started' notification at 2AM. But now that I check Windows>System logs, I see an error this morning a few hours after the backup started showing "The shadow copies of volume D: were aborted because of an IO failure on volume D:" [D = external]. I'll pull the external drive later today from the enclosure and check it for bad sectors. However, that doesn't explain why the server decided to delete all previous backups when the drive was near full, does it?

[Spence]

@link470 - If that external disk is having I/O errors then, really, all bets are off re: the integrity of the data on it. I haven't seen the particular behavior you're seeing w/ Windows Server Backup deleting old backups, and I tend to favor the failing disk as the likely culprit. Given the relative low cost of external disk I would strongly recommend moving to a rotation of at least two external disks. That also helps you get the data off-site, too.

[link470]

@EvanAnderson thanks very much. This is my thought as well. I've read some crazy stories about using Windows Server Backup with 2 disks and WSB getting confused and not backing up correctly, but I may look into the 2 disk option and have one off-site.

[Spence]

@link470 - I have several Customers rotating multiple hard disk drives weekly w/ no ill effects. (Economic acceptance of tape being what it is, this has been a decent alternative.)

[kralyk]

One cheap alternative I've used is RDX drives, but they aren't supported directly within Windows Backup, you need a 3rd party backup utility. I've also setup WinBackup to backup to a local external drive and then do a weekly COPY backup using whatever utility comes with the RDX (to not mess with the archive bit) to the RDX drive then store that offsite for DR. But you have to be willing to give up RPO during a disaster with that setup.