disk auto zero spares

[chriskranz]

Stupid question. But why is there no option to auto-zero disks? I can't see any reason why you would want a disk in the system that isn't zeroed, and the process of zeroing it is always a manual one. Is there a technical reason why there is no option to auto-zero disks?

[adamfox · Accepted solution]

It existed briefly back in the 6.0 days and it was promptly removed after a unusual set of conditions caused an automatic wipe of production data at a site.  It was a pretty rare sequence of events, but it was enough to not auto-zero spares just in case.  The risk even if very small has not been considered to be worth it.

Sorry for the history lesson, but basically it's the law of unintended consequences.

[amiller_1]

Interesting....I often find myself zeroing spares when I'm at customers (notice a non-zeroed disk after "sysconfig -r", mention it to them and then go ahead and zero spares).

[danielpr]

To keep my test bed clean which helps me to create aggregate by keeping all the disks spares are zero'd. But auto zero spares looks a cool option when you test the Storage system often. In my testbed i run a small script to disk zero spares (using CLI). So that next day morning my system will help me to create aggregate without disk zero waiting time.

Thanks

Daniel

[chriskranz]

Cheers for the history lesson, it's good to know the full reasoning behind things.

However if the automated feature would use the same function of "disk zero spares", then surely running in manually has the same inherent risk involved? There's always going to be edge issues (like disk shelves blowing up due to a firmware update), but hopefully that shouldn't stop development? OnTap 6 was 4-5 years ago now?

Along with what Andrew said, I often find customers with un-zeroed spares that they had no idea about. Maybe a disk replacement wasn't zeroed at factory, or perhaps a maintenance recovery didn't zero the disk, but maybe 1 in 3 customers I see have this. Perhaps this is because the majority of customers I see are because they need some guidance, but it is quite a high number.

Or perhaps just an AutoSupport or log message that prompts that their are non-zeroed disks in the system? Similar to when you have un-assigned disks in the system (although this is only on boot).