IRRd-Discuss

[Jason Lixfeld]

Yes, I was able to make the same determination as to the source of the
problem by process of elimination.

Thank you, though.

> -----Original Message-----
> From: Larry J. Blunk [mailto:ljb@merit.edu] 
> Sent: Wednesday, October 09, 2002 2:26 PM
> To: jlixfeld@fastvibe.com
> Cc: irrd-discuss@merit.edu
> Subject: Re: Segfaults after irr_directory
> 
> 
> 
> > Running v2.1.4 on RedHat 7.3 and have found it to segfault.
> > 
> > Oct  8 16:39:01 [1024] UII command irr_directory %s
> > Segmentation fault
> > 
> > The directory is specific and not the default /var/tmp.  If 
> I comment
> > out the directory line, it starts up with no problems but 
> as soon as I
> > put that line back in, it barfs again.  The directory itself hasn't
> > changed at all, the only thing that changed was I removed 
> some of the
> > comments and manually placed config lines in the config to make it a
> > little cleaner looking.  When I rebooted the daemon, it seg'd.
> > 
> > Any ideas?
> > 
> 
> 
>    There's currently a bug in IRRd in the initialization of 
> the IRRD_STATUS
> and <DBNAME>.CURRENTSERIAL files.  Basically, it will not create the
> <DBNAME>.CURRENTSERIAL file if it does not already exist (so you must
> first initialize it by hand).  Further, if you do not pre-create it,
> the IRRD_STATUS file will become corrupted when the registry database
> is first updated.  Can you check if the IRRD_STATUS file 
> exists in your
> irr_directory and send me it's contents?  Also a backtrace of 
> the Segfault
> would be helpful.  
> 
> 
> > gdb ./irrd
> (gdb) run -f irrd.conf
> (gdb) bt
> 
> 
>   You should be able to fix the segfaults by removing the current
> IRRD_STATUS file and initializing the <DBNAME>.CURRENTSERIAL 
> file (put a
> '1' in the first line of the file).
> 
> 
>   Regards,
>    Larry Blunk
>    Merit
> 
>