LOTS of red, changing block count, 100% health?

[another_smith]

Forgive my ignorance, but I am confused by these results. I would appreciate your help in interpreting these... Ran the WRITE + READ on a used 2TB drive I picked up from eBay for cheap. The tests appeared to indicate there were a ton of bad blocks... but then the read test comes back saying just 1 bad block and 100% Health. Is it just that the WRITE test identified all of those blocks as bad and so the READ test didn't even test them? Why is it still showing 100% Health then? At the end of the day, I am trying to confirm whether this disk needs to be sent back to the seller for a replacement. This is my second drive from the seller and the first one was perfect and cleared all tests. Thank you for your time!

Bad blocks.jpg (549.81 KiB) Viewed 6102 times

Bad blocks2.jpg (601.6 KiB) Viewed 6102 times

[hdsentinel]

Thanks for the images and the information about the situation.

This is a quite interesting situation - which can happen and yes, hard to say 100% for sure about the outcome.
Let me explain.

Generally when you saw the lots of red blocks during the write pass, it means that the hard disk could not be properly written.
This *may* be related to the hard disk itself, but in some cases (usually when external enclosures / docking stations used) it may be related to the chipset of the external enclosure / docking station. Some of these chips may overheat during excessive use/data communication and may perform a "reset".
The result is that the hard disk is temporarily unavailable for any software trying to access (Hard Disk Sentinel, to perform the write test) in this case.
It does not automatically mean that all sectors are wrong - just could not be written.

Then, when the read part of the Write+Read test started, the read test could successfully read back these sectors too. This is why their color from red changes to purple, to indicate they are not bad (as they could be read) - but could not be verified to match the written data.
This is why the previously displayed 2863 "bad" blocks removed (as they could be tested) but also the count of "good" blocks do not indicate that all blocks are good - because we can't be 100% sure.


If possible, personally I'd try to connect the drive differently, for example by a different enclosure / docking station (if now used) or directly to a SATA port. This may make things more stable during the test and if the outcome is same, you can be sure that the hard disk is not really perfect.

Also you may try to perform disk test partially on (for example) the second half of the hard disk this way:
- select Disk menu -> Surface test and select the appropriate test type on the disk (for example Write+Read test)
- select Configuration tab in this window and in the "Preferences" section, enable the "Limit testing to specific blocks" option
- specify first block (for example 5000) and leave the last block as 9999, to test only the 2nd half of the hard disk only.

If possible, please use Report menu -> Send test report to developer option. Personally I'd be curious to see the complete status of the drive, to verify error count(s) and possible changes, so the best would be to send developer report before and after the test.

To be honest, personally I'd worry about the very first block, which is suspicious.
Not 100% sure if the drive still could not be written - or just the communication restored exactly in that block (when the write pass finished) and this resulted in the verify errors there.
The very first block is critical in the operation (to store Master Boot Record and so) so it is important that it should be really error-free.

Personally I'd perform additional tests to ensure that the hard disk is really correct or not.