Disk format corrupted disk? - Printable Version
+- IRIX Network Forums (//forums.irixnet.org)
+-- Forum: SGI/MIPS (//forums.irixnet.org/forum-3.html)
+--- Forum: Hardware/Triage/Repair (//forums.irixnet.org/forum-11.html)
+--- Thread: Disk format corrupted disk? (/thread-2861.html)
+- IRIX Network Forums (//forums.irixnet.org)
+-- Forum: SGI/MIPS (//forums.irixnet.org/forum-3.html)
+--- Forum: Hardware/Triage/Repair (//forums.irixnet.org/forum-11.html)
+--- Thread: Disk format corrupted disk? (/thread-2861.html)
Disk format corrupted disk? - synthetix - 05-02-2021
I was trying to wipe a SCSI disk using the fx.64 program, and I chose the [fo]rmat option. This totally locked up the machine, and now when I try to access the disk, the fx program just says "Media error: Media format corrupted, block #0." Anything I try to do, I get this error. The drive was working fine before I tried the format command.
Has anyone encountered this problem? Is there any way to fix this?
RE: Disk format corrupted disk? - weblacky - 05-02-2021
While I've never seen this error and assuming you have access to an Adaptec SCSI controller on a PC, you COULD try to low level format it from within the Adaptec BIOS. It's risky..but at this point you don't care. A controller-issyued low level format will basically reset the block defect list and perform a surface scan, layout tracking marks, and mark bad blocks again. Assuming this isn't a larger symptoms, you could basically attempt to format for bad blocks. However your disk must start normally to do this.
It will either work and exploded, at this point I guess either is acceptable?
Update 1: By the way, manpage claims FORmat shouldn't be done on a hard or fixed disk. It's possible format meant low-level format? A partial or bad low-level format can corrupt fixed disk media if not done 100% correctly.
https://nixdoc.net/man-pages/IRIX/man1/fx.1.html
UPDATE 2 : This confirms it, format MEANS low-level format! In SCSI parlance, destroying the bad block, track layouts, etc... list implies a low-level format.
http://retrogeeks.org/sgi_bookshelves/SGI_Admin/books/IA_DiskFiles/sgi_html/ch02.html#LE27674-PARENT
OK, from your description, I'm confused...you did inadvertently issue a low-level format. However when you say "locked up the machone", how did you observe this and did you shutdown during the operation? Low-level format are done by drive firmware...not the SCSI controller. The controller issues the command then just waits. It's totally normal for nothing to appears on the screen until it's done...like hours! There are no progress bars and may not be any updates on progress until complete.
So if you can describe what you saw you can help others, but you interrupted a low-level format and power-cycled the drive...if that's what you did...then your disk format on disc platters is ruined, period. Tracking markings for the drive head and info were being rewritten when you interrupted it, most drives cannot recover from this as it's done at the factory initially and in the field uses an in-memory info of what they looked like during boot during the reformat. Rebooting removed that info from drive ram and the format didn't complete so it's not on the disk anymore during boot. It's no bricked.
It's gone, you've learned an expensive lesson...FX meant low-level format...yeah they labelled it wrong. But that's what happened, you have a partially formatted disc track layout now, your drive head won't be able to track on disk anymore.
I've done exactly 6 low-level formats in my life. One for bad block rescanning and 5 to change RAIDed SCSI Drives from 520 byte block to 512 byte block (via linux sg-tools). It was nail biting, but I knew I was doing it and even used a UPS. I can't stand FX, I say auto - root or option drive and that's it. So I've never dared press any other option.
RE: Disk format corrupted disk? - synthetix - 05-02-2021
(05-02-2021, 08:07 AM)weblacky Wrote: So if you can describe what you saw you can help others, but you interrupted a low-level format and power-cycled the drive...if that's what you did...then your disk format on disc platters is ruined, period. Tracking markings for the drive head and info were being rewritten when you interrupted it, most drives cannot recover from this as it's done at the factory initially and in the field uses an in-memory info of what they looked like during boot during the reformat. Rebooting removed that info from drive ram and the format didn't complete so it's not on the disk anymore during boot. It's no bricked.
Correct. I power cycled the machine during the [fo]rmat because I (mistakenly) thought the machine was totally locked up. The cursor froze and I didn't hear any drive activity. I should have heeded the warnings in FX that low-level formatting was probably unnecessary.
After reading your post, I tried the low level format on the drive again, but this time I just let the machine sit there, locked up. And guess what? After several hours the cursor came back to life and FX reported the format was successful. I was able to repartition the drive and do a fresh OS install. So disaster averted! Well, the drive was cheap anyway so it wouldn't have been that big of a deal, but I don't like destroying working things unless absolutely necessary.
So thanks for replying! Your good advice made me do what I should have done in the first place!
RE: Disk format corrupted disk? - weblacky - 05-03-2021
Wow, I’m shocked you got a second chance! I guess good for you, glad you were able to get past it.
Cheers!
RE: Disk format corrupted disk? - robespierre - 05-03-2021
If the drive responds to an INQUIRY command, it is generally able to accept a FORMAT UNIT command. The geometry (numbers of tracks and spare sectors) and the primary defect list (the PLIST) are not overwritten by formatting and are not lost. Software is able to report progress during a FORMAT: however, SGI's fx tool does not do so and is unresponsive until the command completes.
All this is quite foreign to (P|S)ATA drives, as their protocol entirely lacks a low-level format facility. It's proprietary vendor tools or nothing.
RE: Disk format corrupted disk? - weblacky - 05-03-2021
I’ve only experienced timer interfaces that progress based on the initial estimate given in response to format command.
I’ve also used drives that were made after 2000 where firmware was part of platter layout (sector -1). If platter access lost so was firmware content. Older drives with ton-o-chips used IC driven firmware storage then this might have worked and been the case. Modern minimal SCSI from early 2000’s held firmware on plater. No platter, no brain during startup which mean never became ready. That’s been my experience. Older drives were more forgiving.
RE: Disk format corrupted disk? - robespierre - 05-03-2021
The FORMAT UNIT command only erases the user data area and has no effect on firmware. If there is a physical problem with the HDA (platters were removed and reinstalled) that may prevent the drive from working.