OS/2, eCS & ArcaOS - Technical > Storage

JFS partition corrupted -- what is the root cause, and how to fix it?

<< < (2/3) > >>

Thomas Hellsén:
Thank you all who have answered so far.

Boris, Eirik: I tried running CHKDSK F: /F:3 from a command window in the management console in the eCS 2.1 installation program, and it states the following, accompanied by a number of short beeps:

The current hard disk drive is: F:
The type of file system for the disk is JFS.
The JFS file system program has been started.
Internal error: devices.c(532): Read error 23 (CRC)
Internal error: devices.c(532): Read error 23 (CRC)
JFS0068: CHKDSK  Superblock is corrupt and cannot be repaired because both
primary and secondary copies are corrupt.  CHKDSK CANNOT CONTINUE.

This does suggest that Doug is correct: the disk may indeed be going bad. It could definitely be worth copying the still-good E: partition and the backup of the F: partition to a new hard disk and see how that works. Then I can let DFSee do a long-time test of the drive and see if it comes out faulty.

Doug: I think I formatted the JFS partitions with eCS 1.2 before the eCS 1.2r update. I have no idea if the actual filesystem data differs between the two eCS versions' JFS implementations, but I haven't heard of any interoperability problems either. Not to say I hear or know much about that, though.
As for clean-marking the partition in DFSee, I assume the partition cannot be read at all (by eCS) since both superblocks are corrupted.

I have a backup that I made before transferring the hard disks to a new system, so most of the files on F: should be okay. It might be worth downloading JRescuer and see if it can do anything with the partition. If it can recover files, it may be worth buying it to recover those files I changed or added since the backup. It will take awhile to get this fixed but I'll post back when I have something new.
/Thomas

eirik:
Hi again,

fortunately, you do have a backup (that hopefully is good).  Your disk issue goes beyond the problems I've encountered before and been able to fix.  A reformat of the disk (preferably from eCS 2.1) will usually render the disk useable again.  Moreover, you may get important feedback to the extent of your problem, in particular if you use the format option where you check for disk integrity.  I am looking forward to hear the feedback from your reformat.

Eirik

ivan:
Hi Thomas,

You didn't sat who the disk manufacturer was so I can't be specific.  It looks as if you need to get the manufacturers disk tools and run them against each disk.  This will tell you if there are problems with the disks and fix bad sectors if possible by replacing them from the reserve sectors.

If they pass the manufacturers tests then I would suggest you reformat using the long format option - just don't try that on a 1.5TB disk unless you have a lot of time before you need to use it. 

Doug Bissett:

--- Quote ---I have no idea if the actual filesystem data differs between the two eCS versions' JFS implementations, but I haven't heard of any interoperability problems either
--- End quote ---

I think the eCS versions are data compatible. It was just the very early OS/2 version that had the problem. I mentioned it because it was a possible cause.

I would suggest that you use your backup (to a different disk, if possible). It will take a long time to try to recover anything from the bad disk. I have found that it is much faster, and easier, to simply recreate whatever gets lost. Of course, if there s something REALLY important, you can try to recover that, by itself.

Those disks are, obviously, IDE disks. You may find it difficult to find replacements, at a reasonable cost. If your new system can use SATA disks, get one of them. Currently, it seems that the smallest, generally, available is about 500 GB. Just don't get one larger than 2 TB, or eCS (currently) can't use it. If your system can't handle SATA (even with an adapter), try the used market, but you take your chances that you will get a bad one. It is usually a waste of time to try to recover a disk that is showing problems, but you may be lucky and a simple, long, format will do it. I wouldn't use that disk for important stuff though.

Alex Taylor:

--- Quote from: Doug Bissett on September 15, 2014, 04:42:24 pm ---
--- Quote ---I have no idea if the actual filesystem data differs between the two eCS versions' JFS implementations, but I haven't heard of any interoperability problems either
--- End quote ---

I think the eCS versions are data compatible. It was just the very early OS/2 version that had the problem. I mentioned it because it was a possible cause.

--- End quote ---

I seem to recall there was one release which changed the filesystem structures slightly, and therefore necessitated a reformat.  I don't remember if it was a late IBM update or one of the BootJFS releases. In any case, reformatting with the latest eCS JFS.IFS and UJFS.DLL in place is probably advisable.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version