$i30 error and now everything is gone :(

[Christoph Mueller]

Hi,

This morning I switched my harddisk from a promise ide controller to
the primary ide controller of my mainboard. (I did handle the hardrive
very carefully and it had no errors before)
when booting windows it said that the disk had to be checked for
errors, and then it displayed cryptic messages for 2 hours. which
could be described as one of the worst 2 hours of my life.
After that all my files were still there, but half of them now contain
only garbage.
i can still access them all, but the content is only random
characters.

The text displayed by chkdsk was like this:


The Signatur of the Multi-sektor-headers for the virtual
Clusternumber (VCN) 0x114 of Index $I30 in the File 0x2b4a
is not correct.
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
Errors in Index $I30 of file 11082 are being repaired.
Index $I30 in file 11082 is getting sorted.
(Translation from german)

This went on for 10 minutes

After that it said "orphaned file FILENAME.EXT is ..." over and over
for 2 hours with lots of filenames. I could not read the whole text
because it went by too fast.

Norton Disc Doctor and chkdsk could not find any errors on the disk
afterwards.

Can anybody tell me what caused this error, and how to avoid it in the
future?

Is there any way to restore the damaged files or at least to find out
which files are corrupt?

Chkdisk displayed a million filenames at startup, but the event log
only contains the first few 100 lines. Where do I find the rest, so I
know which files are affected?


I have a backup of most of the data, only the last month could be
lost, but I need to know how to avoid this error before I can restore,
and I want to know if there is anyway to restore the last month before
overwriting the harddisk with the backup

Thanks a lot for your help

 

[Ndi]

Norton Disc Doctor and chkdsk could not find any errors on the disk

afterwards.

Of course not, they were fixed. Sorta'

Can anybody tell me what caused this error, and how to avoid it in the
future?

Not much. DO test the disk and controller:

* Run chkdsk even if no error is there, ESPECIALLY if you rebooted without
shutting down. NTFS is self-healing but it's not immortal, it can take only
that much damage.
* Get a HDD monitor and monitor SMART status and drive temperature.
* Get a free tool (there are LOTS) and perform a *full* SMART test (takes
some 15 minutes at least). If test doen't come out PERFECTLY, change the
drive. NOW. If you have an IBM disk about 1 year old, there was a massive
manufacturing error and all disks are affected in mechanics. Mine failed
too. BTW, I understand that the HDD division was sold so you might not get
an IBM back (same model, different mark).

Also, be on the watchout for errors. Check the system log periodically
(Event log) and check for signes of malfunction. If NT5 (Windows 2000) gets
to a bad sector it retries, fails, marks it as "bad", LOGS IT and it is
never used again, silently. Shortly, if you don't check the log, faliure is
silent.

Is there any way to restore the damaged files or at least to find out
which files are corrupt?

IMHO, nope.

Chkdisk displayed a million filenames at startup, but the event log
only contains the first few 100 lines. Where do I find the rest, so I
know which files are affected?

Chkdsk logs to a file in the root. I can't remember the name. Upon system
startup, when APIs are available, the log is moved to an event log and the
file is deleted. And since a log is limited to a certain size, all other
data is lost.

You COULD, basically, try to create the file and set permissions so that
the file can be written, but not deleted and thus gaining access to it. But
it's too late now.

I have a backup of most of the data, only the last month could be
lost, but I need to know how to avoid this error before I can restore,
and I want to know if there is anyway to restore the last month before
overwriting the harddisk with the backup

IMHO, no. Restore from backup. But first take up my advice and scan the
disk. Do a SMART test. Install monitors. Then restore and watch how the disk
performs. This time you'll be ready for a solution if one comes up.