How to remove DELAYED sectors

[XENIX]

Hello, recently a friend was gifted to me 5 different old hard drives, and I decided to scan them with the HD REGENERATOR program, but one of them SEA GATE 40 GB turned out to have 3 DELAYED sectors. I tried to fix them by use setting the REGENATE ALL SECTORS option, but still these DELAYED sectors are still there. I tried with other programs like VICTORIA but this program brought out even more errors and defects. Then I also tried with HDAT2, but there the results were that the hard drive is clean without any errors.

After a lot of reading on the internet, for weeks, I came across your forum. I've been trying different options with MHDD for almost a month now, carefully reading the entire manual, and then trying everything step by step with the ERASE, then ERASE DELAYES, then REMAPPING options. I ran another scan with HD REGENERATOR, but again those DELAYED sectors are still there! How can I remove these DELAYED sectors? Please someone with more experience help. Thanks in advance!

PS I specify that, I tested these programs under pure DOS, no Windows! I even start MHDD with BIOS off so it can work directly with the hard disk controller!

[DSfuchs]

What about describing it completely anew?

[DSfuchs]

A sector cannot be delayed and written at the same time. It is best to do this under RAID1 with another disk that is otherwise good but has >= the capacity.

[Arch Stanton]

Ultimately it's the drive that decides if a sector is reallocated, if delay is not severe enough, within parameters and no actual read error, it will stay.

[fzabkar]

The ATA standard has a WRITE UNCORRECTABLE command. This enables the user to designate a particular sector as "pseudo-uncorrectable", irrespective of whether it is actually good or bad. Read commands that access these pseudo-uncorrectable sectors will return UNC errors. Writes to these sectors will clear the pseudo-uncorrectable status and attempt to return the sectors to service if they test OK.

I believe MAKEBAD is one utility that uses this approach:

https://files.hddguru.com/download/Software/Makebad/

[Arch Stanton]

The ATA standard has a WRITE UNCORRECTABLE command. This enables the user to designate a particular sector as "pseudo-uncorrectable", irrespective of whether it is actually good or bad. Read commands that access these pseudo-uncorrectable sectors will return UNC errors. Writes to these sectors will clear the pseudo-uncorrectable status and attempt to return the sectors to service if they test OK.

I believe MAKEBAD is one utility that uses this approach:

https://files.hddguru.com/download/Software/Makebad/

But even then it's the drive that decides if a sector is reallocated, or not?

[fzabkar]

The ATA standard has a WRITE UNCORRECTABLE command. This enables the user to designate a particular sector as "pseudo-uncorrectable", irrespective of whether it is actually good or bad. Read commands that access these pseudo-uncorrectable sectors will return UNC errors. Writes to these sectors will clear the pseudo-uncorrectable status and attempt to return the sectors to service if they test OK.

I believe MAKEBAD is one utility that uses this approach:

https://files.hddguru.com/download/Software/Makebad/

But even then it's the drive that decides if a sector is reallocated, or not?

AIUI, the host cannot force a sector to be reallocated. However, by creating UNC sectors, the OS will add them to $BadClus and take them out of service, if the user is careful in the way that they do this. Presumably a full format would need to be avoided, since this would write to every sector. Instead, a quick format followed by CHKDSK, with a full surface scan, should trap these sectors in $BadClus ... I would think.

In fact, I had a 13GB Seagate drive that had one pending sector. This sector remained pending for the life of the drive because it had been identified as a genuine UNC sector when the file system was built. This meant that the OS never wrote to it.

[Arch Stanton]

I was assuming he meant removing these slow LBA sectors from 'LBA space'.