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

[leafnode-list] Re: Segmentation fault in texpire 20050412

Matthias Andree <matthias.andree@xxxxxx> wrote
in news:m3vf6fsrcn.fsf@xxxxxxxxxxxxxxxxxxxx: 

> "John Carlyle-Clarke"
> <john.cc@xxxxxxxxxxxxxxxx> writes: 
>> I've been having some problems recently with one particular
>> server, where old posts keep getting reloaded as new posts by
>> leafnode.  I've also been seeing messages about damaged groupinfo
>> file, which a fetchnews -f doesn't cure.  I even tried deleting
>> the file in leaf.node for the server in quesion, but no joy.
>> Since I was running an old Alpha version, I decided to update to
>> the current one (20050412).  Having done so, texpire stops with a
>> segmentation fault.  I'm not sure if this is because of the state
>> of my spool or not.

> Looks right, at least this is sufficient to find the segfault bug,
> and it can be caused by a corrupt groupinfo file; although I
> cannot provoke the bug at the moment, it is there. Is your spool
> at the "root" of a mount point, for instance, for a separate news
> partition? Local groups? 

On my setup, /var/spool/news is a symlink to /home/news since my var
partion is not very big, and is quite full.  /home is on its own
partition too.  So, no, my spool is not at the root of a mount
point.  I do have some local groups. 

> What isn't clear to me is if upgrading fixed your fetchnews
> -f/groupinfo problem; if it didn't, please let me know the details
> (log messages in particular), there may be yet another bug to fix.

The messages I was getting were like this:

Apr 18 19:00:00 pc69 fetchnews[16885]: Groupinfo file damaged,
ignoring line: 13 

The line number reported seemed to vary a lot.  As I said, running
fetchnews -f did not fix it. 

> Are there strange files, symlinks, something else in the spool? 

Not that I know of!

> Does the problem persist after an fsck run (you may need to reboot
> into single-user mode to safely run fsck, depending on your
> partitioning)? 

I've run fsck, no problems found.

> This patch should suffice to fix your texpire segfault problem...
> [snip] 

Indeed it does!  Thanks.

I've now not seen the problem (either the messages in the log file,
or the reloading of old messages) since updating to the current
snapshot, but I will monitor things carefully and let you know if I
see anything odd. 

Thanks again!

leafnode-list mailing list