I have to ask a really annoying question to start...

Did you actually put the SAV device in the SAVE pool?

Yes it's in the pool and enabled.

Speaking of size, well that's another issue. I have 300G of free space on the system but when I tried to create the SAVE device of (127000 cyls - same as source device) it said "there aren't enough disks to satisfy the raid/mirror requested". I tried RAID 5 and 2 way mirror to no avail. I then deleted a 127000 cyl RAID 5 device and tried to creae the SAVE device again...same exact error which makes no sense. So just for giggles I created a 2000cyl device and bam it did it but now it says the device is full.

I guess my question should be how to create a SAVE device of 127000 cyls to match the source device? Let me know if you need output from any commands or logs.

This is just a test snap I'm trying to set up.

Would a SNAP session fail to create because the SAV device is too small? Wouldn't that just cause the session to fail sooner? I know the minimum 10% is a rule of thumb, but is there some hard and fast rule regarding minimum size?

Are you creating your SAV devices RAID 5 or RAID 1. If it is RAID 1, that could explain why removing a RAID 5 STD wouldn't give the capacity required to create a RAID 1 SAV dev.

Your RAID 5 device is made up of 127000 cyls spread over 3 drives (+1 for parity)... making the hypers approximately 42333 cyls. For a RAID 1 SAV dev of the same size you would need to be able to create 2 hypers that are 127000 cyls each. You don't have enough space on each individual drive to create the size hypers you need.

Apparently so. I was able to get it going this time and I made two changes.

1. The src device was in "Not-Ready" state so I chose a different one that was in the Ready state.
2. As suggested I bumped up the size to 10000cyls

One last question, how do you guys manage snapshots automatically for doing backups? Say I have a database that I want to snapshot and then kick off a backup (Netbackup for e.g), then terminate the session? If you guys are using SYMCLI and cron jobs how do you ensure the process is atomic?

The only integration I know is PowerSnap .. That blends symsnap and NetWorker together .. Anyone have better/different examples ??

do you have some kind of job scheduler, something more advanced than cron ? ..something like "tivoli workload scheduler" or "control-M" . You can still create bash/korn/perl scripts but use these type of schedulers to create job flows with proper dependencies.