LDAP Issues

[Luke Militello]

Hi all, hopefully someone here can assist me with an LDAP issue.  A 
little background...  I have had LDAP running in a master/slave 
configuration for the past 5 months or so with no problems to report. 
However recently I have had the slave server crash twice and the master 
once.  Parsing through the logs and with help from Nagios, I think I 
have narrowed it down to LDAP (slapd).  The following is an excerpt from 
the slave server syslog file.  After these lines occur, this is when all 
hell breaks lose.

slapd[4608]: connection_read(36): input error=-2 id=2278, closing.
slapd[4608]: conn=2278 op=2 UNBIND
slapd[4608]: daemon: removing 36
slapd[4608]: conn=2278 fd=36 closed

Other things start breaking on the server because the authentication is 
gone.  So bad I cannot even console in as root (which is obviously local 
and not in LDAP); I get a login timeout expired.  The only option is too 
reboot and then all is well for about 5 days or so and the problem 
starts again.

I haven't made any system changes other than a non-deb package I was 
trying to install required some sysv stuff.  Upon installing sysvinit 
and sysvconfig, it hosed a lot of stuff by removing upstart.  Thus I 
reverted with this:

sudo apt-get update && sudo apt-get upgrade && sudo apt-get purge 
sysvinit sysvconfig && sudo apt-get install friendly-recovery 
startup-tasks system-services upstart-logd upstart-compat-sysv upstart 
&& sudo apt-get autoremove

Which from apt/term.log was everything that got removed.  This is 
basically the point when slapd started acting funny.  I don't know if it 
is related but the time lines mesh nicely.

Any help with this is much appreciated.

Thanks,
-Luke

Version: 8.04.2 (Has been on .2 for awhile via proposed repository, so 
it is not related to the problem.)