<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
On 22.10.2009 12:33, Thomas Schmitt wrote:
<blockquote cite="mid:10447175213716@212.46.126.165" type="cite">
<pre wrap="">Hi,
</pre>
<blockquote type="cite">
<pre wrap="">it appears there is a bug in the current cdrkit version that doesn't
allow me to put files bigger than 4GiB on a disk. I tried calling
mkisofs with
-udf -allow-limited-size -iso-level 3
Sadly, no success there, the files on disk remain unreadable. Am I doing
anything wrong here or is it actually a bug?
</pre>
</blockquote>
<pre wrap=""><!---->
Not knowing about genisoimage with large files
and especially not about -allow-limited-size,
i have to stress that in the past at least Linux
was not able to read files >= 4 GiB in ISO images
properly.
(Such files have to be split into several
pieces, called "extents", and the reading system
has to put them together.)
Recent Linux kernels are said to work better.
Mine is of summer 2007 and cannot read some bytes
from the end of multi-extent files in images
created by libisofs. But users of recent Ubuntu
report that there are no problems any more.
man genisoimage says:
-allow-limited-size
...
The result is an inconsistent filesystem and users need
to make sure that they really use UDF rather than ISO9660 driver
to read a such disk.
This sounds slightly frightening.
Maybe you should try without.
If you are using Linux for reading:
Did you enforce mounting of UDF by mount -t udf ?
Have a nice day :)
Thomas
</pre>
</blockquote>
<br>
Actually it appears you guys are right. I just did some testing using a
10GiB file and I ran into no issues at all. The issue I reported was
with a squashfs on an iso actually, so I tested 10GiB squashfs on iso
and that was just fine as well. This magic might be explained by kernel
2.6.31 which came with some fixes that might concern this bug I
experienced. I think for now we can consider this closed, sorry for the
"false alarm". <br>
<br>
If I'm to run into this again, I'll revive this thread.<br>
<br>
-- Sven-Hendrik<br>
</body>
</html>