Improve company productivity with a Business Account.Sign Up

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 993
  • Last Modified:

Secure erase hard drives Dell PE 195

I have a Dell PE 1950 server with a PERC 5/I embedded SAS controller.

I need to securely erase some SAS drives.

I do not see any way in the controller configuration to just pass through the drives, instead I can only create virtual drives, which does not securely erase the drive.

Any ideas?
0
John Tolmachoff
Asked:
John Tolmachoff
  • 5
  • 3
  • 2
  • +1
2 Solutions
 
DavidPresidentCommented:
If you don't want to get a JBOD SAS controller, then configure a RAID1 & erase the volume.  You'll get 99.999% of the drive .. everything except for a few MB of data structures specific to the controller.  No user data will be there.
0
 
andyalderCommented:
If you configure one virtual drive that uses all the space then dban will erase all the user data except for anything that's been relocated to a spare block. Or you can configure each disk as a single virtual drive and do the same erase.

There are some paid-for products that understand the controller so can deconstruct the RAID into individual disks and then print off a certificate with the serial number of the disk on it for "proof" if needed.
0
 
pgm554Commented:
I think if you just set it up as RAID 0 in the controller,you can probably erase that way.
0
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

 
DavidPresidentCommented:
No, do RAID1. It will accomplish the same thing in half the time.   (Either way the metadata will be preserved. So if you have a 300TB disk, would you rather write/verify a 299.9GB LUN, or a 599.8GB  LUN?)
0
 
pgm554Commented:
I think the way that controller works is that it sets it up as 2 jbods when set to raid 0.
From there you just run the secure erase for both at the same time.

Six one way ,half a dozen the other.
0
 
John TolmachoffNetwork AdministratorAuthor Commented:
And somewhere in between is the middle.

;)

Thanks guys, I will try this later today I hope.
0
 
DavidPresidentCommented:
No, it is not the same. If you make it RAID1 then it will take less time. The RAID controller will  be writing the data to the 2nd disk transparently in the background while you write in the foreground.  Your app only needs to write half the amount of data in the foreground.

Since this won't be a compliant erase anyway, then consider the additional benefits of the last phase, the verification phase.  You'll get read load balancing, So if this is the standard DoD write 0, write 1, write random, read, then the last phase will take half as long.

Not only that, but all writes will take less time in addition to having fewer foreground writes.  Why? Because the write cache guarantees 100% of writes will be cached to 2nd drive in RAID1.

But in 2xRAID0 then 0% of the writes will be cached, plus you have to do 2X the number of writes.  [just make sure write-back cache is enabled, which is probably the default]

If you were doing more drives, then RAID1 would be even faster because you wouldn't have to deal with bus saturation as much either.

Since the disks are rather small, since I expect they are pretty old, then why not do a test?

Benchmark total time for 2 instances of doing the RAID0, then do the RAID1.  But make sure that before benching the RAID1 you make sure the controller has finished initializing/mirroring the array before you start.  That will take a few hours at least.

Then we'll see ;)
0
 
andyalderCommented:
Don't want read load balancing after erasing, you would be verifying that half the data on one disk and half on the other has been zeroed out.
0
 
DavidPresidentCommented:
it won't be compliant anyway, unless you have a true non-RAID controller. Since you are writing the data immediately before the validation  phase, then only way a RAID-based config  wouldn't be correct is if you have unrecoverable write errors.

Any unrecoverable write errors on a secure erase mean that the HDD has to be physically destroyed.  So as long as the controller doesn't report any such errors then the data will match.    (and if you get an unrecoverable read error on a disk, then no worries, because you already know the data was erased on those blocks, so you are OK)

Now you do risk that the disks were short stroked (reprogrammed to show that there are fewer blocks on the drive then there really are.  The easy way to figure out if the disks were reprogrammed is to look at the make/model info and read up the specs, then look at the size of the disks that are reported to the BIOS (which should match), or the size of a RAID 1 reported by the O/S .. the unaccounted for data is metadata, and should be around 32 MB).
0
 
John TolmachoffNetwork AdministratorAuthor Commented:
Test A: Created one RAID1 volume, initialized, then did a DoD 5220.22-M wipe, then ran a consistancy check. 2 hours 13 minutes

Test B: Created one RAID0 volume, initalized, then did a DoD 5220.22-M wipe, no consitnacy check available. 3 hours 51 minutes

Test C: Created 2 RAID0 volumes (one for each drive,) initalized then did a DOD 5220.22-4 wipe, no consitancy check available. 4 hours 2 mintues
0
 
John TolmachoffNetwork AdministratorAuthor Commented:
Thanks guys. Not the ideal solution, but it works, since I do not have to certify them, just secure wipe them before scrapping them.
0
 
DavidPresidentCommented:
Like I wrote ... it will do it in half the time, and you even penalized the RAID1 by making it do a 5th pass by forcing a manual consistency check ;)
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

  • 5
  • 3
  • 2
  • +1
Tackle projects and never again get stuck behind a technical roadblock.
Join Now