[torqueusers] Re: pam_authuser segfaults
mischa.salle at uam.es
Thu Feb 8 03:09:24 MST 2007
Concerning the perl scripts, I have found one problem with the epilogue
script, which is a problem for at least two different perl versions
(5.8.0 and 5.8.8). This is what happens:
- after reading in the accessfile the filepointer is at the end.
- truncate ( ACCESSFILE,0) then fills the old file with \000, but
doesn't reset the filepointer, that only happens after closing or with
an explicit seek() statement.
- hence print ACCESSFILE writes at the former end.
At some point one of our authuser files was more than 1GB, filled with
A simple seek( ACCESSFILE,0,0); after reading in the file and before (or
after) the truncate is enough.
It has now been running for about a week without problems...
On Tue, Jan 30, 2007 at 09:56:56AM -0700, Garrick Staples wrote:
> On Tue, Jan 30, 2007 at 05:45:08PM +0100, Mischa Salle alleged:
> > Hi,
> > I've been going through the pam_authuser.c code, and apart from the
> > original segfaults which I had found, there were some more issues which
> > I have (hopefully) fixed.
> > Attached is a new version, it would be great if you could go through it
> > to check for obvious errors. It seems to work fine on our system.
> Thanks for sending it in. I'll check it out this afternoon.
> Has anyone fixed the perl scripts yet?
> torqueusers mailing list
> torqueusers at supercluster.org
Dpto. de Física Teórica C-XI mischa.salle at uam.es
Universidad Autónoma de Madrid Tel. +34 91 497 4897
Cantoblanco, 28049 Madrid Fax +34 91 497 3936
SPAIN Room 506
__ .. ... _._. .... ._ ... ._ ._.. ._.. .._..
More information about the torqueusers