RE: Index corruption after multiple updates

["Howard Chu" <hyc@highlandsun.com>]

Redhat 7.2 has been available for quite a long time already. Questions
regarding availability of updates are best researched at www.redhat.com, not
on an OpenLDAP mailing list. Whether or not disabling threads is an option
for you depends on what components you need. slurpd will not build without
threads, but otherwise it may be a good approach.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support

> -----Original Message-----
> From: Paul M Fleming [mailto:pfleming@siumed.edu]
> Sent: Wednesday, March 13, 2002 2:17 PM
> To: Howard Chu
> Cc: openldap-software@OpenLDAP.org
> Subject: Re: Index corruption after multiple updates
>
>
> Any suggestions?? Disable threads? An alternate thread package? Has
> RedHat released any updates??
>
> Thanks in advance
>
> Paul
>
> Howard Chu wrote:
> >
> > RedHat 7.1 is known to have a broken threads library and cannot
> be safely
> > used.
> >
> >   -- Howard Chu
> >   Chief Architect, Symas Corp.       Director, Highland Sun
> >   http://www.symas.com               http://highlandsun.com/hyc
> >   Symas: Premier OpenSource Development and Support
> >
> > > -----Original Message-----
> > > From: owner-openldap-software@OpenLDAP.org
> > > [mailto:owner-openldap-software@OpenLDAP.org]On Behalf Of
> Paul M Fleming
> > > Sent: Wednesday, March 13, 2002 8:20 AM
> > > To: openldap-software@OpenLDAP.org
> > > Subject: Index corruption after multiple updates
> > >
> > >
> > > I have a definate problem with index corruption under 2.0.23. The
> > > problem can be solved in the short run by doing a slapcat removing the
> > > DB files and doing a slapadd -- the problem usually reoccurs within 24
> > > hours. (commenting out the affected index also fixes the
> problem but (of
> > > course) has a big performance impact). Running slapindex does NOT fix
> > > the problem. This problem is occuring on the master and on the slaves.
> > > The version of _all_ software on each is identical (we run cfengine so
> > > I'm sure of this) The version information is as follows:
> > >
> > >       Linux 2.4.2-2
> > >       Red Hat Linux release 7.1 (Seawolf)
> > >       openssl-0.9.6
> > >       db3-3.1.17
> > >       openldap-2.0.23
> > >       cyrus-sasl-1.5.24
> > >       krb5-libs-1.2.2
> > >
> > >       configure information
> > >               --enable-wrappers \
> > >                 --enable-cleartext \
> > >                 --with-cyrus-sasl \
> > >                 --enable-spasswd \
> > >                 --enable-kpasswd \
> > >                 --with-kerberos=k5 \
> > >                 --enable-ipv6=no \
> > >                 --with-tls \
> > >                 --prefix=/usr
> > >
> > > Are there any known issues with db 3.1.17?? I couldn't find any open
> > > bugs at http://www.openldap.org/its/
> > > Could this be thread or concurrency related?? I haven't built a test
> > > harness yet to be able to reproduce the problem, but that shouldn't be
> > > too hard.
> > >
> > > Suggestions???
> > >
> > > Thanks
> > >
> > > Paul Fleming
> > > SIU School of Medicine
> > > Springfield Il USA
> > > pfleming@siumed.edu