Re: [Sks-devel] Segfault in v1.0.4

[Yaron M. Minsky]

Is there anything in log.db or log.recon after you run this?

Anyway, it does sound quite plausible that the problem could be bitrot
somewhere in the libraries that SKS depends on.  I certainly don't see
what else could be causing these problems.  Maybe a recompile would
help. 

By the way, what were the contents of sksconf for this run?  And was
there anything in the log files?  The lines you sent me do not look like
complete logs, since SKS among other things writes its version number in
the log before anything else.

y

On Sat, 2003-11-01 at 23:40, Daniel Johnson wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Discovery is using a hair over 26mb right now.
> 
> I've had sks_db, sks_recon, build, pbuild, and (I think)
> clean_keydb  give segfaults so far.  Earlier, I could run build
> and sks_db one (1) time without trouble.  If I shutdown and
> restart sks_db it would blow up.  pbuild never would work. 
> After that, my only choice was to delete the database
> directories and try again.  
> 
> address@hidden:~$ sks_db -debuglevel 10
> 2003-11-01 22:14:19 Membership: <ADDR_INET
> 68.91.150.38:11370>(dannyj.dynip.com 11370)
> Segmentation fault
> 
> I also tried that with no 'membership' file.  It just leaves off
> the first line.
> 
> address@hidden:~$ sks_recon -debuglevel 10
> Segmentation fault
> 
> I killed and rebuilt the database, which got it working one more
> time.  Added a key via the web interface.  Shut it down, tried
> to restart.  Segfault.  Tried to delete/rebuild again.  Didn't
> work this time.
> 
> I'm reminded that ocaml did not compile properly the first time.
>  Following their instructions, I re-ran the compile and it
> worked.  Do you think a total recompile would help?
> 
> Or maybe I'm overlooking the obvious: This system is old, and in
> dire need of a complete rebuild.  Perhaps some of the libraries
> are a bit quirky after all this time.  At least the kernel is
> recent (2.4.21).
> 
> Yaron: If it would help, I can arrange for you to have access to
> that account.
> 
> On 1 Nov 2003 at 21:47, Yaron M. Minsky wrote:
> > Hmm.  This is quite strange.  The sks_db and sks_recon
> processes do NOT
> > need huge amounts of memory --- on the order of 10-30 each. 
> Right now
> > they're together taking 25 megs on my machine.
> > 
> > Are you talking about sks_db and sks_recon or
> build/fastbuild/pbuild? 
> > If so, exactly which one fails, and why?  keyserver.bu.edu is
> a 128 meg
> > machine, and SKS works just dandy on it.
> > 
> > So, I guess the next step is to provide some more detailed
> info about
> > exactly which executable fails and where.  By the way, the way
> to get
> > more debugging info is not to set the debug flag (which is set
> by
> > default anyway) but to set the debuglevel to something high. 
> The
> > default is 3, 5-6 is pretty verbose, and 10 will give you a
> printout of
> > every message passed through the system.
> > 
> > y
> 
> - -- 
> Through the modem, off the server, over the T1, past the
> frame-relay,
> < < NOTHIN' BUT NET > >
>  
> Daniel Johnson
> address@hidden
> http://dannyj.come.to/
> Public PGP Keys & other info: http://dannyj.come.to/pgp/
> 
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (MingW32) - GPGshell v2.95
> 
> iD8DBQE/pIqT6vGcUBY+ge8RAiDzAJ9sukKMm9o7yhM9WQYl6d4xmRYbtACgx1Ll
> CO/NhorqkGRlCop5rXQa2r4=
> =1xTX
> -----END PGP SIGNATURE-----
> 
> 
> 
> _______________________________________________
> Sks-devel mailing list
> address@hidden
> http://mail.nongnu.org/mailman/listinfo/sks-devel
-- 
|--------/            Yaron M. Minsky              \--------|
|--------\ http://www.cs.cornell.edu/home/yminsky/ /--------|

Open PGP --- KeyID B1FFD916
Fingerprint: 5BF6 83E1 0CE3 1043 95D8 F8D5 9F12 B3A9 B1FF D916