[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Running apt*, help needed



Bob Proulx wrote the following on 12/06/2011 09:32 PM:
Dennis Wicks wrote:
Just to clarify, the "bad" system is only "bad" because whein I was
installing an update, for Open Office if memory serves, aptitude
crashed and left "something" in an unusable state that causes
segmentation faults and relocation errors whenever I try to run
anything "new".

Yes.  But a couple of points.  One is that OpenOffice.org has a very
large cone of dependencies behind it.  The problem could have been one
of the libraries pulled in because of that cone.  The top of the cone
might be simply OpenOffice.org and very small but the bottom of the
cone might be very large!

That is, the things that were running when aptitude crashed are
still running. Like Thunderbird, Iceweasel, a couple x-window
terminal sessions, and Opera. Anything I try to start now fails to
do so for the aforementioned reasons.

But along the way it was suggested that you use a chroot and use the
dpkg --root option and other things that required you shutting the
system down and mounting the disk on another system.  So with that in
mind there won't be any process that are still running.  All of them
would be new processes..

As far as I can tell the disk is physically OK. But I will check it
out, just in case.  I'll also check the lock file, etc.

I haven't seen anything that makes me think the disk has failed.
Probably okay.

If everything you run is segfaulting the the problem is probably in
one of the shared libraries that has a large cone of dependents.  But
the number of really critical shared libraries is fairly small.

  $ ldd -d -r /usr/bin/apt-get
  	linux-vdso.so.1 =>  (0x00007fff787ff000)
	libapt-pkg.so.4.10 => /usr/lib/libapt-pkg.so.4.10 (0x00007f314355c000)
	libutil.so.1 => /lib/x86_64-linux-gnu/libutil.so.1 (0x00007f3143358000)
	libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f3143054000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f3142dd2000)
	libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f3142bbb000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f3142837000)
	libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f3142633000)
	libz.so.1 => /usr/lib/libz.so.1 (0x00007f314241b000)
	/lib64/ld-linux-x86-64.so.2 (0x00007f3143a93000)

That is a fairly manageable number of libraries.  I would check the
md5sum of those against the known good list.

Bob

Bob, et al,

Oops! Another thing I don't know how to do: Mount a disk from a system that isn't running! When I say "mount" I meant something like "mount //host/share " where fstab defines to mount it on "/some/local/mount/point". Obviously not what you thought I meant!

I didn't know that "shutting the system down and mounting the disk on another system" was implied in the recommendation to use "dpkg --root". That will mean that I will have no means of communication with the outside world until I get it fixed!

I'll have to think about that for awhile! I don't think I have the room to mount another hard drive in my operational system. I'll probably wind up having to use a live CD of some flavor. Any preferences??

Thanks!
Dennis


Reply to: