Does anyone know how you can zero out a specific disk(s) ? The "disk zero spares" command or "spares_zero" diagnostic command seem to zero out ALL spare disks only.
Thanks, George
Hi,
Le mardi 27 juillet 2004 à 09:53, George Kahler écrivait:
Does anyone know how you can zero out a specific disk(s) ? The "disk zero spares" command or "spares_zero" diagnostic command seem to zero out ALL spare disks only.
I never saw a command that do only one spare disk.
If you have an FC card for a windows/linux box you can use seatools to format a drive all by itself. But then you also must have a spare disk shelf or a little cable adapter.
-tavis.
On Tue, Jul 27, 2004 at 09:53:13AM -0400, George Kahler wrote:
Does anyone know how you can zero out a specific disk(s) ? The "disk zero spares" command or "spares_zero" diagnostic command seem to zero out ALL spare disks only.
Thanks, George
George Kahler wrote:
Does anyone know how you can zero out a specific disk(s) ? The "disk zero spares" command or "spares_zero" diagnostic command seem to zero out ALL spare disks only.
Hello George
If your focus is on pre-zeroing:
disk zero spares "only" zeros all non zeroed disks.
If you just want to zero one selected single disk start: vol add vol0 -n -d YOUR.DISK
He will check wether the disk is zeroed, and if not do so ... After zeroing, he will detect the '-n' option and therefore not add it to the volume. It will stay a spare disk. :-) This is very useful for older ONTAP releases who don't have the "disk zero spares" command.
If your focus is on data destruction:
If you really want to wipe the old data for data security reasons you should buy a "sanitation" license from NetApp to have a legally approved wipe out process. disk sanitize start ...
If you want to do it the "cheap" way: Fill and destroy the volume at least 5 (up to 10) times. Those data reconstruction services can tell by seeing if your "1" is hanging at 90,91 or 97.5% wether the history of this bit is 11011 or 01101 or 11001 or ... => To fill it use random data (by using the network clients as your random data source with a lot of network trafic :-( ). Using "hammer" inside the filer is definitely not safe enough. => Buy the sanitation license if you don't want to get killed by your colleges.
Smile & regards Dirk
Dirk.Schmiedt@munich.netsurf.de (Dirk Schmiedt) writes
[...]
If you just want to zero one selected single disk start: vol add vol0 -n -d YOUR.DISK
He will check wether the disk is zeroed, and if not do so ... After zeroing, he will detect the '-n' option and therefore not add it to the volume. It will stay a spare disk. :-) This is very useful for older ONTAP releases who don't have the "disk zero spares" command.
Neat trick! I'll add that one to the collection...
If your focus is on data destruction:
If you really want to wipe the old data for data security reasons you should buy a "sanitation" license from NetApp to have a legally approved wipe out process. disk sanitize start ...
"Legally approved" ?
The description in the man page doesn't sound all _that_ strong:
| If no patterns are specified, the default is 3 using pattern 0x55 | on the first pass, 0xaa on the second, and 0x3c on the third.
which is roughly what a Unix "format analyze" will do (though better than a simple zeroing, agreed). No deliberate write-with-track-offset, for example.
If you want to do it the "cheap" way: Fill and destroy the volume at least 5 (up to 10) times. Those data reconstruction services can tell by seeing if your "1" is hanging at 90,91 or 97.5% wether the history of this bit is 11011 or 01101 or 11001 or ... => To fill it use random data (by using the network clients as your random data source with a lot of network trafic :-( ). Using "hammer" inside the filer is definitely not safe enough. => Buy the sanitation license if you don't want to get killed by your colleges.
Colleagues? (I presume). It's the litigious users I would be worried about, not my colleagues...
[Or perhaps you really did mean "colleges". Working as I do for a collegiate University, I am certainly aware of the animosities that arise between the Colleges and the University. But I don't think they have actually got around to organising a hit squad yet ...]
Chris Thompson University of Cambridge Computing Service, Email: cet1@ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QH, Phone: +44 1223 334715 United Kingdom.
Hello Chris, hi Toasters
If your focus is on data destruction:
If you really want to wipe the old data for data security reasons you should buy a "sanitation" license from NetApp to have a legally approved wipe out process. disk sanitize start ...
"Legally approved" ?
"Almost" legally approved: The Sysadmin Storage Management Guide pdf, Chapter 5) Disk Management, p.233 informs us: NetApp ONTAP disk sanitation is compliant to U.S. DoD and DoE security requirements. You, as an inhabitant of the "51st state of america" ;-), and probably native english speaker can interpret this better than I can do.
The description in the man page doesn't sound all _that_ strong:
| If no patterns are specified, the default is 3 using pattern 0x55 | on the first pass, 0xaa on the second, and 0x3c on the third.
The manpage and the Storage MG differ at ONTAP 6.5.1R1: The Storage Management Guide pdf says: 6 cycles with these three pattern are the default behaviour. The manpage of "disk" talks about just one single default cycle with those three different patttern which definitely would be not be enough.
I could not check the correctness of the value in the Storage MG yet, because these activities are invisible at the "statit" disk analysis, and the disk sanitation only shows the progress in % not in cycles.
=> If somebody is really interested in the (current) truth: Please measure the time needed to wipe it with -c 1, -c 2, ... and without any "-c X" to find out the real (current) default number of cycles. :-) I didn't have the time/need to do so yet.
Or maybe just a netapp engineer can clarify the default cycle value?
Colleagues? (I presume). It's the litigious users I would be worried about, not my colleagues...
You guessed right. I meant colleagues. (I should not be sitting at the keyboard after twelve hours of giving NetApp lectures in a foreign language.) b.t.w. If you flood the network unnecessary, they all will try to kill you (imho). Even if they are not your colleagues. :-) Even I would do so. ;-)
Smile & regards! Dirk
Dirk Schmiedt (and others) wrote:
The description in the man page doesn't sound all _that_ strong:
| If no patterns are specified, the default is 3 using pattern 0x55 | on the first pass, 0xaa on the second, and 0x3c on the third.
The manpage and the Storage MG differ at ONTAP 6.5.1R1: The Storage Management Guide pdf says: 6 cycles with these three pattern are the default behaviour. The manpage of "disk" talks about just one single default cycle with those three different patttern which definitely would be not be enough.
b.t.w. The manpage (default = 1 cycle) is correct, the Storage Management Guide isn't. (ONTAP 6.5.1R1) => Use the max. value (-c 7) for disk sanitiation explicitely.
disk sanitize start 8a.18 disk sanitize start -c 1 8a.19 disk sanitize start -c 2 8a.20 disk sanitize start -c 4 8a.21 disk sanitize start -c 7 8a.22
disk sanitize status
sanitization for 8a.18 is 50 % complete sanitization for 8a.19 is 50 % complete sanitization for 8a.20 is 28 % complete sanitization for 8a.21 is 15 % complete sanitization for 8a.22 is 8 % complete
-
Aurélien Beaujean -
Chris Thompson -
Dirk Schmiedt -
George Kahler -
Tavis Gustafson