thanks for linking my initial posting to this thread. Somehow I couldn't find a suitable topic, but this thread has it all. I'm really scratching my chin after reading this thread (and several linked ones). Chaining is - mostly - the devil.
Btw, why are you trying to run only a single job at a time?
Mostly out of fear that lets say 10 starting backup jobs will immediately create 10 snapshots that will exist longer than in a serial approach in total, maybe filling up the VM dastastore.
Follow-up question: When is a snapshot actually created? At around the job start, or when the ressource manager releases ressources to a backup job, allowing it to "actually" start?
A typical recommendation is to start all the jobs simultaneously (or with a short intervals) and limit the maximum number of proxy/repository concurrent tasks, if the resources load is your concern. This will allow you to reduce the backup time even more.
Indeed. I'll have to check this out. Starting jobs on Saturday will honor the Saturday AF flag for this job, even if it is delayed into Sunday because of ressource availability?
Thanks so far.