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

Re: Bug#221706: segfault when reading package lists



Matt Zimmerman wrote:
On Wed, Nov 19, 2003 at 11:20:14AM -0800, Ken Treis wrote:


Matt Zimmerman wrote:

- One of the upgraded packages (wget) seems to have changed to using UTF-8
in its control file.  Though, I upgraded to the same version on an
i386/unstable machine and had no problems.

Not sure if this is the right approach, but I tried `dpkg -P wget` and the problem remains.


Can you send a gzipped copy of your entire status file?


I am not sure what (exactly) I did to fix it, but I'm no longer experiencing this problem. The problem was solved through some combination of:

* Selectively upgrading packages to match a known good system
* Installing `file`
* Unmounting /dev/shm (originally a tmpfs filesystem)
* Rebooting

For the record, I began doing selective upgrades first. After each upgrade, I'd re-run `apt-get update`, and in all cases the segmentation fault persisted.

I then played with my sources.list and discovered that only two package lists triggered the problem: sarge/main and security sarge/updates non-free. This seemed especially odd, since the "security sarge/updates non-free" package list is empty.

At one point, I noticed that the problem machine had /dev/shm mounted. So I unmounted it. At about the same time, I also noticed that the known good machine had `file` installed whereas the problem machine didn't. So I installed the `file` package. After that, scanning of the empty sarge/updates non-free package list no longer caused segmentation faults. However, the sarge/main package still tripped it up.

At that point, I finally became brave enough to reboot the system (remotely; it's on the other side of the state from me). After reboot, apt was able to scan the sarge/main package with no problems.

It's possible that the reboot did it, or it's possible that this machine has a bad piece of RAM and a reboot relocated things sufficiently. I really don't know.

I'm just glad that the system is working now. I'm sorry the problem wasn't more reproducible, but I wanted to pass on as much information as I could in case any of was helpful.

--
Ken Treis
Miriam Technologies, Inc.



Reply to: