NVME wipe takes a long while

[DaanSelen]

Is it normal for a wipe with method DoD 7 passes to take like 14 hours on an NVME drive?

Its starts off with a transfer speed of 1000-2500 MB/s and then later drops to 80-100 MB/s

Is that normal?

[PartialVolume]

Is it normal for a wipe with method DoD 7 passes to take like 14 hours on an NVME drive?

Its starts off with a transfer speed of 1000-2500 MB/s and then later drops to 80-100 MB/s

Is that normal?

It's normal for there to be an initial burst of speed. This is because nwipe uses cached I/O. The writes are initially filling the CPU RAM with writes to disc many many times faster than the NVME disc can write but once the memory used for cached I/O has filled up, the speed drops to the true speed of the drive. This bust only lasts a few seconds but also depends on how much RAM your system has.

80-100 MB/S is slow for NVME. Something is causing a bottleneck. What's the hardware configuration. Is the NVME drive connected by M.2 on the motherboard or a NVME USB adapter? What make/model of NVME? What's the temperature of the drive as shown in nwipe?

[PartialVolume]

Allright. That does looks pretty clear. I assume nwipe does not do this mechanism yet. But its planned?

Yes, don't ask me when though. The plan being to integrate it so a ATA secure/enhanced erase is as easy as doing a normal wipe complete with progress.

[DaanSelen]

Allright. That does looks pretty clear. I assume nwipe does not do this mechanism yet. But its planned?

Yes, don't ask me when though. The plan being to integrate it so a ATA secure/enhanced erase is as easy as doing a normal wipe complete with progress.

That sounds clear. I would love to see that.

[PartialVolume]

I'd love to know the status of both projects. Since we at our company want to use this for our customers. I want to know I do it well. Right now we just do a Department of Defense 7 passes on EVERY disk. Sata-SSD, Sata HDD, NVME drive - EVERYTHING. And I wanted to check if I am not screwing over those drives.

Google over provisioning of SSDs. There is a chunk of memory that can't be written to by just writing to blocks, hence why a ATA secure erase should be done on a SSD. The secure erase should in theory (but apparently not always) clear the over provisioned area of memory.

If you don't do that, somebody with the necessary skills and equipment could remove the memory chips from the board and put them into a reader which allows them to see all memory, that over provisioned area of RAM could contain sensitive data. That's why you should use ATA secure erase. PRNG should also be applied because it's been found that some SSDs firmware was very bad at erasing all data. I don't remember the name of the US university that did the study but it's linked to in other posts.

[DaanSelen]

I'd love to know the status of both projects. Since we at our company want to use this for our customers. I want to know I do it well. Right now we just do a Department of Defense 7 passes on EVERY disk. Sata-SSD, Sata HDD, NVME drive - EVERYTHING. And I wanted to check if I am not screwing over those drives.

Google over provisioning of SSDs. There is a chunk of memory that can't be written to by just writing to blocks, hence why a ATA secure erase should be done on a SSD. The secure erase should in theory (but apparently not always) clear the over provisioned area of memory.

If you don't do that, somebody with the necessary skills and equipment could remove the memory chips from the board and put them into a reader which allows them to see all memory, that over provisioned area of RAM could contain sensitive data. That's why you should use ATA secure erase. PRNG should also be applied because it's been found that some SSDs firmware was very bad at erasing all data. I don't remember the name of the US university that did the study but it's linked to in other posts.

And this memory is non-volatile I assume due to its capability to hold data?

[DaanSelen]

I've learned a lot about SSD's the last few hours. Our old method also just did DoD 7 passes... which also was not sufficient then:
https://blog.securaze.com/why-traditional-sanitization-methods-as-specified-in-dod-5220.22-m-are-not-sufficient-for-ssd-erasure-and-can-leave-recoverable-data-behind