Corrupted files udf/ext2
Hallo, [was thread: readerror, no access rights] i made some more trys, copying the linux2.4.18 source-tree (about 140MB) using packetwriting to cdrw. Kernel 2.4.19-pre10 with packet 2.4.19-4, TEAC-Writer W524E IDE, used with SCSI-Device-Emulation. Write cache is disabled in the kernel. I used the same cdrw, reformatting it accordingly to ext2 and udf. I copied 2times to ext2 and 2times to udf, each after a fresh format of the cdrw. After writing to cdrw i copied back the date with cp -a to harddisk, and then made a diff from this copy on hd to the original. With ext2: No diffs in both copys With udf: diffs in both copys, but not in the same files. The number of different files was about 15 to 20. So in the moment i believe, the problem is not in packetwriting, but in udf-format. Would it be possible, to try other formats, like dos-fat, on cdrw? BTW: We had this in the other thread: CDRW's with ext2 can't be read in my Cdrom (Toshiba 6402B). Is that a problem of this drive? Greetings-- Email:KOALASOFT@GMX.DE
to see if it is udf or packet i tryied making a filesystem with udf format. Then i copyied the source and then compared. The results: No errors, so i think this is a packet problem, not udf On Fri, 2002-06-21 at 19:00, ulrich mensfeld wrote:
Hallo,
[was thread: readerror, no access rights]
i made some more trys, copying the linux2.4.18 source-tree (about 140MB) using packetwriting to cdrw.
Kernel 2.4.19-pre10 with packet 2.4.19-4, TEAC-Writer W524E IDE, used with SCSI-Device-Emulation. Write cache is disabled in the kernel.
I used the same cdrw, reformatting it accordingly to ext2 and udf. I copied 2times to ext2 and 2times to udf, each after a fresh format of the cdrw.
After writing to cdrw i copied back the date with cp -a to harddisk, and then made a diff from this copy on hd to the original.
With ext2: No diffs in both copys With udf: diffs in both copys, but not in the same files. The number of different files was about 15 to 20.
So in the moment i believe, the problem is not in packetwriting, but in udf-format. Would it be possible, to try other formats, like dos-fat, on cdrw?
BTW: We had this in the other thread: CDRW's with ext2 can't be read in my Cdrom (Toshiba 6402B). Is that a problem of this drive?
Greetings--
Email:KOALASOFT@GMX.DE
-- To unsubscribe, e-mail: packet-writing-unsubscribe@suse.com For additional commands, e-mail: packet-writing-help@suse.com
Hallo Mario Medina Nussbaum ,
to see if it is udf or packet i tryied making a filesystem with udf format. Then i copyied the source and then compared.
The results: No errors, so i think this is a packet problem, not udf
so it seems, there is all possible and nothing regular. Peter has experienced coruppted files on ext2, i only on udf (since using 2.4.19), and you have no problems with udf. Maybe it could be a problem, related to certain CD-writers?
With ext2: No diffs in both copys With udf: diffs in both copys, but not in the same files. The number of different files was about 15 to 20.
Greetings-- Email:KOALASOFT@GMX.DE
No, i think you missunderstand me.. my english isn't very good. Writing CD-RW with UDF HAS errors for me. What I was done to test UDF is making this: dd if=/dev/zero of=udf.iso bs=1024k count=650 mkudf udf.iso mount -o loop udf.iso path cp -aR /usr/src/linux-2.4.18 path diff --brief -r /usr/src/linux-2.4.18 path/linux-2.4.18 May be it is more clear now :) If i do the same, but using my cdwriter, then i get corrupted files, like everyone. On Fri, 2002-06-21 at 19:00, ulrich mensfeld wrote:
Hallo Mario Medina Nussbaum ,
to see if it is udf or packet i tryied making a filesystem with udf format. Then i copyied the source and then compared.
The results: No errors, so i think this is a packet problem, not udf
so it seems, there is all possible and nothing regular. Peter has experienced coruppted files on ext2, i only on udf (since using 2.4.19), and you have no problems with udf.
Maybe it could be a problem, related to certain CD-writers?
With ext2: No diffs in both copys With udf: diffs in both copys, but not in the same files. The number of different files was about 15 to 20.
Greetings--
Email:KOALASOFT@GMX.DE
-- To unsubscribe, e-mail: packet-writing-unsubscribe@suse.com For additional commands, e-mail: packet-writing-help@suse.com
ulrich mensfeld
Hallo,
[was thread: readerror, no access rights]
i made some more trys, copying the linux2.4.18 source-tree (about 140MB) using packetwriting to cdrw.
Kernel 2.4.19-pre10 with packet 2.4.19-4, TEAC-Writer W524E IDE, used with SCSI-Device-Emulation. Write cache is disabled in the kernel.
I used the same cdrw, reformatting it accordingly to ext2 and udf. I copied 2times to ext2 and 2times to udf, each after a fresh format of the cdrw.
After writing to cdrw i copied back the date with cp -a to harddisk, and then made a diff from this copy on hd to the original.
With ext2: No diffs in both copys With udf: diffs in both copys, but not in the same files. The number of different files was about 15 to 20.
So in the moment i believe, the problem is not in packetwriting, but in udf-format. Would it be possible, to try other formats, like dos-fat, on cdrw?
BTW: We had this in the other thread: CDRW's with ext2 can't be read in my Cdrom (Toshiba 6402B). Is that a problem of this drive?
Hi; Im running 2.4.18 + packet-2.4.18 patch. After initializing the cdrw, putting ext2 on it, and mounting it as a packet device (/dev/pktcdvd0 on /mnt), I tried this: cd /usr/src tar cpSf - 2.4.18.virgin | (cd /mnt ; tar xvpSf -) diff -r /usr/src/2.4.18.virgin /mnt/2.4.18.virgin (thats 157M kernel src) diff found two corrupted files. cdrw is a 8/4/32x 'PHILIPS ' 'PCRW804' ata drive using sg driver. Paul set@pobox.com
Hi, Am Sam, 2002-06-22 um 02.00 schrieb ulrich mensfeld:
Hallo,
[was thread: readerror, no access rights]
i made some more trys, copying the linux2.4.18 source-tree (about 140MB) using packetwriting to cdrw.
Kernel 2.4.19-pre10 with packet 2.4.19-4, TEAC-Writer W524E IDE, used with SCSI-Device-Emulation. Write cache is disabled in the kernel.
I am using Kernel 2.4.19-rc1 with packet 2.4.19-pre4 patch, Yamaha Writer CDR4416S. I copied the complete kerneltree (about 180 MBytes) to a UDF formatted disk. When I compared theese files, I found 5 different files. The errors are always the same: first 2048 bytes (1 block) are 0 (binary zeroes), the rest is ok. Btw, copying a large number of small files is really terrible slow. Manfred
participants (4)
-
koalasoft＠gmx.de -
Manfred Kreisl -
Mario Medina Nussbaum -
Paul