Replies: 6 comments 2 replies
-
Yes
Yes
No, not a problem. Some users just like to see the disc full of zeros rather than random data as random data could be mistakenly seen as encrypted data while zeros is an empty disc when looked at in a hex editor.
Yes, that's correct.
A PRNG stream is created from a single seed, if you know the seed you can reproduce the stream. That's how a prng verification works, nwipe knows the seed it uses to create the stream, therefore it can recreate the exact same prng stream for verification that it created for the wipe . Also the seed is randomly generated so is different for every wipe and for every disc being wiped
Personally, I think it's very important because a PRNG and verification really does a good test of the disc and I like to makesure the disc is working correctly. I've seen discs that will pass a zeros wipe but fail a ones or prng wipe because of a stuck bit at 1 in the drives electronics, so personally I prefer a prng wipe as it provides me with confidence that the drive has no faults.
Depends, what ever method you use I would make sure a verification occurs on a prng pass. So for me I typically would do a prng with blanking, verification on all passes, so in all the disc is written twice and read twice. But that's just me. Others may prefer other methods.
No, neither on a HDD or SDD for verification, however the wipe may be a different matter on a SSD, ATA secure erase may be preferable on a SSD, or at least keeping a traditional wipe on a SSD to a single PRNG pass plus ATA secure erase.
Yes, in fact this should be the method you use if wiping very small drives. If a drive, typically an old USB memory stick is smaller than your CPU RAM you should always verify as a separate operation having removed the USB device and plugged it back in. So if your CPU RAM is 16GB any USB stick smaller than that should be verified with a zeros or ones as a separate operation. The reason being is that nwipe uses cached I/O so if you verify a disc that's smaller than your CPU RAM nwipe will actually verify the cache not read the disc. If the disc is larger than your RAM then the disc gets read. This will not be necessary when nwipe gets switched to Direct I/O when it controls buffering rather the kernel. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you so much for the great answers, they give me confidence that what I'm doing is correct.
I also have other questions, which are not related to this topic, I hope it is okay to ask them here (I just created a GitHub account today, so I don't know the rules). I have read another issue post, in which you mentioned that to erase an SSD you would do an ATA Secure Erase using hdparm, followed by a verify ones or verify zeros method. Thank you in advance for the help :) |
Beta Was this translation helpful? Give feedback.
-
|
@trikti Sorry, I must have missed your last post back in March, I'll close this issue but if you still would like me to answer those questions let me know. |
Beta Was this translation helpful? Give feedback.
-
|
I have another question - not sure if it was fully covered above. DBAN used Verification to actually fill the disk with zeros as a last pass - this fact is described in the "Verify Last Pass" description to this day: If that's not changed, in the default scenario we have a situation, where after the erasure (set in "Method") is completed, the disk is filled with zeros twice:
I'm not sure if that's the case in ShredOS, because I'm just starting migration from DBAN, but will be very happy to hear from the author about this. Also I don't understand, why in the "Method" we have "Verify Zeros" and "Verify Ones" - that should be placed somewhere else, as those variants do not erase data and are used just for verification. It is not a good idea to mix such things, especially in security software. |
Beta Was this translation helpful? Give feedback.
-
|
Also if you switch blanking on, this writes zeros (0x00) only. Nothing else. Why have a blanking pass? Well sometimes when wiping discs users may prefer to see a disc full of zeros on erasure rather than data patterns or random data, which can look like encrypted data. It really depends on personal preference or local policy. |
Beta Was this translation helpful? Give feedback.
-
|
So in summary, unlike DBAN nwipe/ShredOS separates blanking and verification. Blanking is optional so you can just do a zeros write and verify. Verification in ShredOS doesn't force a final zeros blanking pass. Verify last will verify whatever happens to be the last pass in the method. If you switch blanking on which will be a zeros pass, then verify last will verify the zeros blanking pass. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I would really like to understand how ShredOS works, especially the concepts of blanking and verification. I'm not very experienced so I apologize if some of my questions are basic or obvious. I just want to make sure I fully grasp how these processes work.
If I understand correctly, blanking simply means filling the entire disk with zeros.
Additionally, I’d like to understand how verification works.
3) If the previous round wrote only 1s or only 0s, verification checks whether the disk contain the expected values (only 1s or 0s). But what happens if the previous round used a PRNG stream? In that case, does verification still make sense and what exactly does it check?
4) How important is verification?
5) Does verifying between passes matter, or is the verification after the final pass the only one that truly matters?
6) Does verification causes additional wear on the disk?
7) If I didn’t enable the last pass verification during the wipe, would running 'Verify Zeros' (or 'Verify Ones') afterward achieve the same result?
8) Would it be sufficient to disable verification and manually check the hex data myself to confirm that the wipe was successful?
Thank you
Beta Was this translation helpful? Give feedback.
All reactions