I'm just checking how other people run their backup jobs to check whether I'm doing it right.
I'm the admin of a small company, we have a server with vmware running some windows 2008 servers which I backup with vRanger and Backup Exec. Our backup server is a physical windows 2003 machine with a few TB diskspace.
What I am doing currently is like this :
The full backup jobs are kept for 5 weeks before being overwritten, So 5 weeks on disk and 5 weeks on tape.
Each month I make one tape and intend to keep those for 2 yrs.
The incremental jobs run for mo till thu and are being kept for two weeks :
For some new software application I had to create another daily backup job (because of that job having to run at a specific time) so because of that I created another set of backup-to-disk folders and seperate job.
So with all of that and some other backup-to-disk folder I already have over 40 backup-to-disk folders!!
I like splitting backup jobs, I like to have seperate job running for seperate servers. It gives me a clear overview of the job and whenever a job failed for some reason I can start that job again which the only backup what I want.
If I was to have one backup job which does all my servers, and it fails then I cannot just start it again.
Anyhoo, Is this a normal approach, is anyone else doing is like this?
I am using BackupExec 2010R2 with the AVVI agent to back up my 2 ESX servers. I also have SQL agents and Exchange agent to do GRT backups of SQL and Exchange.
I run Exchange backups as a completely seperate job straight to disk and then duplicate that to my SAN. The SAN replicates it offsite.
For everything else I run full backups to disk that duplicates to tape on Saturday. There are 3 individual jobs that append to the disk and tape media. The 3 individual jobs are: 1) Physical servers and CIFS data (NetApp filer) that are backed up using the standard RAWS, 2) VMware Servers w/o SQL using AVVI directly from the SAN with GRT enabled for File level restores, and 3) VMware Servers w/ SQL again using AVVI with GRT enabled for SQL and File. This way I get appropriate granular backups and restores.
Then Monday-Friday I run differential backups of the same 3 jobs doing disk-to-disk-to-tape again.
All of the jobs are handled through BackupExec policies and applied to selection lists.
I also have SAN to SAN replication that tapes place as a completely separate part of my business continuity plan.
It was a bit of a pain to get everything working properly, but now I'm fairly pleased with the results and mostly glad that I didn't buy a separate backup product for file level backups and another product for VMware backups.
You seem to have implied that you are backing up the "Backup Exec" backup to disk folders via the file system - this is not recommended you should be duplicating the backup sets in the backup to disk folders, NOT backing up the backup to disk folder themselves
The reason for this recommendatuion is
1) a duplicate backup can restore directly to the original source instead of needing to be restored to the b2d and then restore to the source.
2) bkf file names can get re-used - if they have been re-used and you then restore an older version you can introduce inconsistencies between the BKF file content, the Backup Exec Database Content (BEDB) and the Catalogs content which can make the restore process very difficult.
Note above comment applies to the Backup Exec disk folders - the vRanger ones may not have the same problems.
BTW have you investigated the Backup Exec AVVI option to backup the VM's instead of using vRanger?