The only low-budget test ideas I have:
1) Start rsync, and while it is running unmount the source drive,
either by software or simply by physically disconnecting it. But I am
not sure whether this results in the same I/O endless timeout, or
wether this causes different error signaling than a bad block.
2) Connect the hard disk via an open cable with a certain "security
distance" to the computer and harddisk, and then apply a strong
magnetic field on the cable, to cause transfer data corruption, which
might be intercepted by CRC, possibly causing a I/O error, or certain
retry loops, being similar to the scenario which you may want to test.
3) Program some virtual /dev/ices, and put I/O errors in there
knowingly.
The additional interesting question, which arises out of this topic
for me:
How can one voluntarily "smash" only a particular sector of a hard
disk drive? Even if you had a test HD for "smashing", you don't know
whether you will completely destroy the HD or only partially, and if,
which sectors/parts. The only idea I have so far, is to write a file
knowingly to a certain sector with a low level tool (which one?), and
then somehow get the disk firmware to mark one or some of the desired
sectors as "bad" (but again, with which tool? And how without that a
"healing mechanism" tries to move the affected file to a healthy
block?).
The CD scratching a la Tomas Gustavsson seems the only easily
achievable solution. But then it is not sure whether the OS does the
reading retries or whether the optical disk drive itself retries
reading.
Am 22.12.2009 um 13:18 schrieb Tomas Gustavsson:
Yeah, I got an Input/Output error when running strace. I don't have
the luxury to smash my harddrive so testing with a CD is my only
choice (afaik) right now. Still, I do think that rsync should give
up after a long time, but it doesn't. So, any advice?
2009/12/22 Paul Slootman <paul+rs...@wurtel.net>
I suspect that this is more down to the OS and the CDROM drive both
trying very hard to read the damaged data. You would have to do a
strace
or similar to determine whether rsync is actually getting IO errors
from
the OS or not.
On Tue 22 Dec 2009, Tomas Gustavsson wrote:
So I took a CD (which I scratched with a needle) and mounted it to
the file
system. There after I started the backup job which went on forever
and never
got completed. It seems that rsync refused to understand that the
file it
tried to copy was located on damaged sectors on the CD, and there
fore was
unreadable.
--
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html
