sys$etc:RZTOOLS.EXE

[Wim Van den Wyngaert]

What is the format command doing that the vms command INIT/ERASE is not doing ? Should I always use it before using (old) disks ?

If I have disks with bad blocks, does it help to reformat them with rztools or is e.g. delete/erase doing enough ?

Wim

[Joseph Huber_1]

RZtools does a SCSI-level format, i.e. (re-)writes SCSI block headers and other low-level data.

The VMS INIT does nothing at SCSI level, it merily writes data to an already SCSI-formatted disk (namely the ODSx structures), and the /ERASE switch writes a pattern over all free (data-)blocks.

RZtools is more like ANALYZE/MEDIA/EXERCISE in overriding the disk completely with patterns, and registering bad blocks.

[Wim Van den Wyngaert]

So, if I do the analyze the rztools are not needed ?

BTW : why is it unsupported ?

Wim

[Karl Rohwedder]

Wim,

to support it, it needs testing with all sorts of SCSI widgets and devices and so on, and resource are limited, I suppose...


regards Kalle

[Wim Van den Wyngaert]

Also. If init/erase or del/erase is done, bad blocks should be put on the bad block list of the drive. I already noticed that this is not always working (e.g. stations).

Does this means that the only way to get the bad blocks on the list is to use the rztools ? Why should I re-init the scsi stuff ?

Wim (confused)

[Joseph Huber_1]

In my opinion, the only reason to use a SCSI format: if a disk develops many new bad blocks, then as a last resort trying to see if it is cured after formatting. If still bad ==> dumpster.

[Wim Van den Wyngaert]

Still confused : does this means that putting bad block aside doesn't work ? (by the disk I mean)

Wim

[Joseph Huber_1]

No, I think bad block recognition usually works (inside the disk driver or explicitly via analyze/media).
But bad block recognition does not make a disk better, it just marks them (and replaces them from a reserve list).
At the time when the reserved block list is exhausted, formatting MAY or MAY NOT help.
(At least once I had such a disk: got rather often bad blocks, no more since formatting).

Therefore I said formatting as last resort before throwing the disk away.

[Wim Van den Wyngaert]

How can I verify if the bad block list is exhausted and init at scsi is needed ?

In my experience with 100+ alpha stations, bad blocks are in about 30% of the cases recovered by the disk itself or by re-read (and some luck). In all other cases I need to fix the bad block in a not used file and mark it non-move.

Still confused

Wim

[John Gillings]

RZTOOLS is an unsupported utility for poking at low level SCSI stuff.

There are no production level operations that NEED it (otherwise there would be a supported tool). It's unsupported because engineering resources are better spent maintaining and enhancing utilities that people really need.

It's there to give access to the low levels to those who know and care about it. It's more for the hobbiest/tinkerer than anyone else. It can also be used by support people to confirm a diagnosis of a failed SCSI device, or possibly update firmware. If you need to perform any of these operations, you will be given detailed instructions by your support engineer.

If you have a disk that you suspect has exhausted its bad block replacement pool, and you care about your data, then it's a very false economy to attempt to "fix" it by doing a low level format. After all, those blocks were bad once, what makes you think attempting to reuse them will suddenly make them better?

Any suspect disk should be replaced by a newer, larger, more reliable and cheaper drive.