There must be a better way than what I'm doing. In my opinion what I'm doing is overly complicated, the backups take far too long, and I'm running out of resources. Input on what I could do different to reduce the chaos is much appreciated.
I have Backup Exec 2014 console on Windows 2008 server with a Tandberg storageloader for DDS1 tapes which hold 720GB each. It backs up 9 Windows 2008 servers. I use the method backup to disk then to tape once a week for keeping something offsite. Of those servers, 2 are backed up in full 5 nights a week, 2 are backed up full once a week and incrementals 4 nights a week, and 5 are backed up in full once a week. Generally the backup to tape is performed once a week on Sundays but there’s not enough time so I separated the verifications to run on Mondays then take tapes home Monday nights. To get all 9 servers on tape now takes four tapes.
For back to disks I have two Synology NAS for most data (highest priority of course) and four lower end external drives for the rest.
No jobs have dependencies meaning I have to carefully calculate the time it will take to run the backups and schedule accordingly. I must check my jobs often and adjust to accommodate increases in data. I was hoping the move to version 2014 would help with the backup speeds but it didn’t. Instead the backups to disk are a little slower which adds to the frustration.
Attached is a screenshot of my calendar which shows the crazy complexity of my backup schedule. It’s out of control. When I recently upgraded from version 2012 to 2014 all those jobs had to be recreated. I’ve proposed backup to cloud at least once a week to replace the tapes but it may not happen until next year’s budget.
I wish I could do the backup to tape and verifications in one job that covered all servers.
I wish I could create dependencies between server jobs to recoup the time lost between jobs. In other words, when server 1’s full backup ends, server 2’s full backup starts.
I don’t like doing incrementals because it will take longer for a full restore. Meaning, I would need to start with restore of the last full backup then restore each incremental backup one a a time until current. I have two servers with incrementals because I don’t have enough time on the schedule for doing more full backups for those servers. I don’t believe changing from incremental to deduplication will be an improvement. It might save space and time on the backup side of things but not on the full restore end which matters.
I don’t think I can improve upon the fact I need to get at least one full backup to tape of each server weekly and then to put that weekly full backup to tape. Maybe it’s time I added a second Backup Exec console, Tandberg storageloaders, and dozens of more tapes so I could run two jobs simultaneously?