anyone else having mediacheck problems with FC3t1 images?
John Reiser
jreiser at BitWagon.com
Mon Jul 19 22:51:35 UTC 2004
> <3>Buffer I/O error on device hdd, logical block 161345
> <4>hdd: media error (bad sector): status=0x51 { DriveReady SeekComplete
> Error }
> <4>hdd: media error (bad sector): errir=0x30
> <4>end_request: I/O error, dev hdd, sector 1290776
>
> There are several more of these type messages each with the logical
> block and sector incremented by one.
Assume this is FC3-test1-i386-disc1.iso, whose length is 660889600 bytes.
[If it is really -disc2.iso, with length 669511680 bytes, then the
entire rest of _this_ message DOES NOT APPLY.]
Look at this section of "od -c -Ax FC3-test1-i386-disc1.iso | more",
which occurs within the first couple screenfuls:
-----
008370 360 001 \0 I S O M D 5 S U M =
008380 d 0 2 d d a d 1 2 8 4 b b 9 1 4
008390 3 f 2 9 2 c 6 4 0 e b 4 b 6 b 8
0083a0 ; S K I P S E C T O R S = 1
0083b0 5 ; R H L I S O S T A T U S = 0
0083c0 ; T H I S I S N O T T H E
0083d0 S A M E A S R U N N I N G
0083e0 M D 5 S U M O N T H I S
0083f0 I S O ! !
-----
Notice the "SKIPSECTORS = 15;" which [probably] means that the last 15
<somethings> (either IDE 512-byte disk sectors, or 2K-byte CD-ROM physical
blocks) should not be included in the _internal_ md5sum (which is not
the same as the md5sum of the whole .iso file). This is from the "-pad"
commandline argument to mkisofs.
From the last console message: Assume "sector 1290776" means "IDE 512-byte
sector 1290776". Then 1290776 corresponds to byte 1290776 * 512 =
660,877,312, which is 12,288 bytes short of the end of the .iso file.
That 12,288 bytes is within the 15 * 2048 byte -pad region. So probably
that particular error can be ignored, on the basis that the error arose
by reading beyond the physical end of data, when mediacheck was reading
30 * 2K bytes at a time as an efficiency hack. The blocks that correspond
to the data that the installer will use, should be OK.
I don't know how to interpret "<3>Buffer I/O error on device hdd,
logical block 161345" which may be from much earlier, or even from
a different CD-ROM.
--
More information about the fedora-test-list
mailing list