cancel
Showing results for 
Search instead for 
Did you mean: 

Consequences of incremental backups

cf_ch
Not applicable

Hello there,

I've got a question about incremental backups with backup exec (version 12 to 15). We use this kind of backup for our customer which backup their servers on a nas over iSCSI. Once a week (for example on monday) a full backup job is scheduled. On the other days an incremental backup job is planed. So now in fact I have two questions:

- What happens to the restore possibilities when an incremental backup fails on tuesday, but runs fine again on wednesday? Are we able to use the backup from wednesday and afterwards or is the whole backupchain damaged?

- What happens to the restore possibilities when the full backup fails on monday but the incremental backups afterwards are running fine? What happens then to our backupchain?

We had in the past some troubles with backup exec incremental backups. Some times on a restore process it says "backup files are missing" because incremental backups failed. But the mentioned files in fact were present on the storage, so I want to make sure our configuration of incremental backups is fine... :)

Sincerely,

Claudio

1 REPLY 1

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

The usual recommendation (based on being cautious rather than the exact effect) is that if a backup in an incremental sequence (either a full or an incremental) has failed, run a full backup as soon as possible to make a new chain as the status of archive bits, change journals & CBT etc cannot be guaranteed

 

It is kind of difficult to answer your question in depth as the answer depends on what you were backing up, when the failure happened AND what the failure actually was.

For instance if you are just backing up files and folders then you can still do partial restores without the complete chain back to the full, so not running a new full might only lose a very small amount of data. However if it is a GRT backup of a database or a VM then you need the full chain to mount the image against to restore (except the sets that were created before the failure which will still work)

 

Similarly for Vmware VM backups if the snapshot process is successful then this probably resets CBT. However if the backup of the snapshot then fails if you just run another incremental it could be based on the wrong CBT status and split the CBT chain within the incrementals (and a restore might be corrupt) Note this is a hypohetical thought of a possible consequence as I am not sure exactly when CBT is reset in the VMware backup process.

 

However if the failure was missed job due to scheduling issues then it won't be a problem as the chain will not be affected.