splay synthetic fulls over more days

Wow, that might be what I’m looking for. I was contemplating just manualy deleting the auto-created oodles of “nearly identical” auto synth schedules (one per VMgroup). 

On the other hand, going forward that could cause confusion for such clients where the auto schedule works just fine.

It was suggested to me by a vaulter maybe a long time ago to abandon synth fulls so that fulls could just rip blocks and nearly complete the backup in as much time as the incrementals run (with the obvious additional SAN load), but, this synth full business running 3500 jobs without much control sometimes leaves the MA really jammed up for a day at a time, subsequently the aux copies are jammed up for the next day because of huge TB’s to process vs scattering them throughout the week.

The other thing I noticed is that in the console, viewing “all” schedules for say the month, you can’t see “any” synth full jobs scheduled - so again, no way to accurately predict when they are going to fire off.

IMO I think this is steering me to use that regkey and just know that fulls will be “ours to schedule” regardless of type. I can see where a small CommCell would take advantage of auto-synth full schedules, but the bigger the number of VMs it becomes what we have which is rush-hour traffic jams. 

These sorts of discussions are what I miss at the old GO conferences :(

thanks

Also if you could  share the sizing information of your environment specific to the VM backups ? i.e the no. of VSA’s you have and the No. of MA’s in your environment which are used for the VM backups . Assuming you are using SAN transport mode here ? Are you also using Intellisnap specially for the high transactional VM’s ? I hope the environment has been sized appropriately based on the current load of FTE . 

If you are only using streaming backups , Abandoning SF’s entirely may not be a viable solution .Staggering the schedules is also recommended based on your backup window . 

The Regkey recommended above is also a good suggestion . 

Hi,

We are doing more with less as they say, 1 MA/VSA combo and all SAN transport.

No intellisnap - tried it, but it’s technically dev, so timing not as important.

As I sort of eluded to, pre-plan days I micro-managed all subclients as needed to distribute and control load, same number of VMs, same everything in fact, and things were very dialed in. The move to plans has made things more challenging mainly for the synth full aspects. I think I’m going to try the key and custom schedule via derivative plans as I don’t really want to revert totally away from plans since I’m 95% there.

thanks