Fixdisk help

[/df]

...For the last 2-3 years the Reallocated count has been 8, until this new set of 16 appeared in 197/8. Those original 8 were successfully moved by fixdisk, so it seems odd that this bug has now surfaced unless it's due to a change in fixdisk.

No, it does seem that you had 8 originally reallocated, then another 8, and 16 pending/uncorrectable that either couldn't be or haven't been reallocated. As it's an advanced format disk with physical sector size 4k, or 8 logical sectors, what this actually means is 1, then another 1, and 2 pending/uncorrectable physical sectors. Even so, you may well be right that the disk could be just on the edge of decline and to treat the current symptoms as a hint to replace.

...I'll start working toward a new disk instead. It's going to need it sooner or later.
...

So after replacing the disk, copy any precious contents to it from the old one, then just overwrite the old disk, which may well clear the pending/uncorrectable sectors. Then you can use the old disk as a backup or whatever. If you didn't already have a USB-SATA bridge you could order up a USB enclosure in which the old disk could reside.

Not that I'm suggesting it, but the paranoid can get this 3.5" RAID enclosure which can hold 2 x 2.5" x 9.5mm disks in (among others) a mirrored configuration. For HDR use you'd need some way of making the status LED visible from the front panel; maybe there's room next to the USB socket to drill a hole for a light pipe.

 

[MikeSh]

precious contents

It's SWMBO's machine. 90% is stuff recorded long ago - a year or more in many cases - and not watched yet (there's about 150 films). So there won't be much of that precious stuff.

No, it does seem that you had 8 originally reallocated, then another 8, and 16 pending/uncorrectable that either couldn't be or haven't been reallocated.

I had 8 ancient reallocated plus the 16 pending. At some point around the first fixdisk run that became 16 + 16. That suggests to me it managed to reallocate 8, but another 8 pending appeared.
Is that right (or plausible)?

 

[/df]

It's SWMBO's machine. 90% is stuff recorded long ago - a year or more in many cases - and not watched yet (there's about 150 films). So there won't be much of that precious stuff.

Rather you than me to explain why the 150 films that were being saved aren't there to watch during lockdown ...

I had 8 ancient reallocated plus the 16 pending. At some point around the first fixdisk run that became 16 + 16. That suggests to me it managed to reallocate 8, but another 8 pending appeared.
Is that right (or plausible)?

I would say so. And anyway ... But it's only 4 physical sectors out of 250M; in Full HD terms (which I know verges on apples v oranges) that's 1/30 of a pixel.

 

[MikeSh]

Rather you than me to explain why the 150 films that were being saved aren't there to watch during lockdown ...

So far we're not watching any more than before, but I guess that might change.

But it's only 4 physical sectors out of 250M; in Full HD terms (which I know verges on apples v oranges) that's 1/30 of a pixel

That's an interesting way of putting it. But people have said in the past that even a few can cause problems if it's in a critical area.

I've ordered a Seagate Pipeline now, so I may think about shifting at least some of her stuff across. Not sure what hardware I've got - have to research the various copy options.

 

[MikeSh]

Well today there was a disk alert in WebIf and the 16 in each of lines 5, 197 & 198 have all turned into 24.

I thought 5 could only increase if you ran fixdisk and it moved them from 197/8. I guess that idea is wrong.

 

[MartinLiddle]

I thought 5 could only increase if you ran fixdisk and it moved them from 197/8. I guess that idea is wrong.

The idea is definitely wrong; the reallocated sector count can increase independently.

 

[/df]

... because the disk firmware can relocate a marginally bad sector if a successful read occurs or an unreadable sector that is being overwritten.

And

... But people have said in the past that even a few can cause problems if it's in a critical area.
..

Sure, and the best solution is to have the disk try to repair itself from scratch, without having to worry about preserving existing data or whether the repair will finish in time for some important recording (it won't).

 

[MikeSh]

No alerts today, but 197/8 have dropped from 24 to 8, but 5 is still on 24.
I guess it's being clever in the background.

 

[/df]

Param #5 should never decrease (once relocated, always relocated).

Apparently the disk firmware has decided that two previously suspect physical sectors have completed their parole and should be allowed back into the general pool, without a stain on their character, etc. I hope this will be shown to have been "clever".