Another NFS nightmare

ME dugan at passwall.com
Sun Mar 24 19:23:25 PST 2002


On Sun, 24 Mar 2002, Lincoln Peters wrote:
> When I do that, lsof hangs.  It produces no output, even if I don't grep it.

At this point, you may want to examine creating a new exported system
root.

Did you use tar when you made the first one? There are, and have been bugs
in tar for handling certain dev special files, and some flags on some
files as well as symlinks. I found cp -aR much more effective in this
arenawhen I set up my new network booting roots. Also another useful tool
is symlinks (had to install it separate) to convert all aboslute symlinks
into relative ones. (Useful when you are modifying files in the exported
root whithout first chrooting to it so that you dont accidentally end up 
modifying the server's config files.)

The exported system roots I have made "just worked" for booting the
system. Certain applications did not work due to the export being "ro",
but when "rw" the single station worked fine with memory being the only
issue. (Setting up the "ro" was a big pain for 80+ stations to share the
same exported root.)

If you make a separate root env, then you can even compile your kernels in
this env, and be certain everything is "there" that is needed. I do most
of my package updates for the exported trees in chrooted env. The problems
you report should not be occuring.

As yet another test, if you have your exported root on a separate
partition, and the "root" of that partition, then you could look into
making another lilo entry with a kernel from this area and that kernel's
"root" being that actual partition and see what happens from the server
when you boot up. I suspect you will see the same problems. Perhaps some
files are missing on your exported tree.

If/when you make a new exported root, you can compare the number of files
in that root to the root from where you copied looking for size and
numbers of files to make sure they were all copied. A simple du -s for the
root copied and the destinate should offer a very close to 2:1 ratio.

-ME

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS/CM$/IT$/LS$/S/O$ !d--(++) !s !a+++(-----) C++$(++++) U++++$(+$) P+$>+++ 
L+++$(++) E W+++$(+) N+ o K w+$>++>+++ O-@ M+$ V-$>- !PS !PE Y+ !PGP
t at -(++) 5+@ X@ R- tv- b++ DI+++ D+ G--@ e+>++>++++ h(++)>+ r*>? z?
------END GEEK CODE BLOCK------
decode: http://www.ebb.org/ungeek/ about: http://www.geekcode.com/geek.html



More information about the talk mailing list