Chaining Jobs and Job Start Date

[Dark-Sider]

Hi,

in what most of you probably would consider a small environment I have 7 Backup jobs of mixed sources (VMs, Windows Agents) witch each job containing several VMs or computers.

I know, that the repository settings allow to set the max amount of concurrent jobs on that repo. However I like micro-managing performance and therefore all jobs are chained to each other so the NAS where the repo is locted serves the best performance for each job, and has some wiggle room to also serve the reads for a tape job.

I was wondering, why some settings on my jobs fail to execute (like do a active or synthetic full backup every Saturday). Then I realized most jobs actually don't run on Saturdays since the active full backup on Saturday of my first job already lasts well into Sunday.

My solution is now to limit concurrent tasks on the repo to 2 and start each job parallel to each other and let the scheduler decide when to pause / unpause the jobs. I would assume this solves the issue active / synthetics are skipped.

However in my opinion the preferred solution would be that all chained jobs inherit the job date from the first job in the chain, thus eliminating any issues when jobs run past midnight.

thanks,
Dark-Sider

[david.domask]

Hi Dark-Sider,

Yes, this behavior is noted in the User Guide, and while I understand your idea, this is why we generally do not recommend job chaining. Active / Synthetic fulls are meant to give full backups on the defined schedule. Your solution is the correct one with controlling the concurrent tasks.

While I understand your idea, it would also mean that it becomes much more difficult to predict and control when full backups are performed, and especially for Active Fulls any unexpected delays in the backup may result in production machines doing an Active Full during production hours unexpectedly, which may be quite undesirable depending on the workload running on the machines.