Ok, a follow up on server backups and what appears to be happening. These are my observations only and they are not intended to be any sort of answer
to the WHS 2011 Great Backup Mystery. A little verbose but hopefully will be of interest to someone.
The backup disk I used is an internal WD 80GB.
This is too small for normal use obviously, but very suitable for experimenting and enabling multiple backups to be run in a reasonable time frame.
The first backup consisted of the WHS system drive plus a couple of small shared folders giving a total of 21GB. Once the backup had completed the
dashboard HD tab showed the correct space used/space free.
I then added a further 13GB of data and started another backup. Looking at the HD tab showed that the backup had calculated the space required and
free space prior to commencing the data transfer. The figures had changed slightly at the end of the backup but not by much. After backup 33.5GB used and 41.1GB free.
Added a further 17 GB and re ran back up.
As the preparing for backup was running the used space on the backup disk was calculated to be 50.9GB. Once the backup had completed this dropped to 50.1GB free. A corresponding change was also shown in the free space.
At this point, noted something I was not expecting.
The total data being back up, as shown in the HD tab was 58.7 GB. The space used on the backup drive was only 50.9GB.
It appears that the backup routine had identified identical files in stored in different locations and only backed up one copy!. When I added data
for the second backup, the 13 GB of files copied over consisted of 3 copies of the same files put into different folders. The data usage figures are not an exact match but close enough to make me suspicious that the backup routine is checking CRC's or such
like and if finding a file match just adding a pointer in the backup database. Do not know with any certainty and would suspect it does not do any such checking but thought the discrepancy should be mentioned.
Right back to the continuing backups. I ran another backup without adding any extra data. There was no change in the disk used/disk free.
Added a further 20,5GB which will be enough to pretty much fill the backup drive.
After backup the used space was 69.7Gb and free space 4.8GB.
The preparing for backup calculation had given figures of used 70.6GB and space 3.9GB.
Added a further 4.3 GB and re ran backup. Leaving 0.4GB free on the backup drive.
Crunch time, copied a further 2GB to a folder and ran backup expecting it to fail as the data I had added was cumulative and nothing had been deleted.
This is precisely what occurred.
The more I thought about the backups, given what I had done above, there was no way the backup routine could free up any space by deleting a backup.
Each backup only added the differences in data, so the accumulating data was effectively in every backup set. The backup routine could not delete anything as there was nothing unique in the first backup that was not included in the last successful backup.
It was time to redo the above but in the first backup have some data that I can delete prior to running backup 2.
Repeated the above steps with the exception, that before running
the second backup, deleted several GB of files. Running the last backup which should have filled the disk, noted the free space increased. The backup completed successfully. The deleted files from the first backup appeared to be
being removed off the backup disk. Unfortunately the backup itself was still listed in the server backups.
To summarise, as with any backup media when you run out of space you need another disk or second drive to do part of the back on.
If, over the period of the backups, data has been deleted off the server it appears, but is not proven, that the backup process does remove the
data from the backup disk.
The above is my take on what occurred during my backup experiment. The figures given are accurate but the conclusions are my own and should be taken
as such.
Dave
The Frog on the Lilypad at Home
