why won't ldap start?

[vaneck]

ok something is really screwed up now.

I just got finished working trying to get my MRTG graphs fixed. Then some of our users where having trouble logging into IMAP through SSL (it would time out and report to the client that SSL was not running when it was). So I went to the admin console and enabled clear text logins while I tried to figure out why ssl was messed up. I did this for the IMAP settings both in the global config and server config. Then I restarted zimbra and I get this message.

Host *.*.*
Starting ldap...Done.
FAILED
db_recover: Finding last valid log LSN: file: 1 offset 5488091
db_recover: Recovery starting from [1][5487967]
db_recover: Recovery complete at Thu May 18 14:45:47 2006
db_recover: Maximum transaction ID 80000016 Recovery checkpoint [1][5488091]
ERROR - failed to start slapd

What I try to run zmcontrol start this is what /var/log/zimbra.log shows:

May 18 14:54:02 lima zimbramon[3608]: 3608:info: Starting services
May 18 14:54:04 lima slapd[3674]: @(#) $OpenLDAP: slapd 2.3.20 (Mar 9 2006 15:10:13) $ root@build.liquidsys.com:/hom
e/build/p4/main/ThirdParty/openldap/openldap-2.3.20/servers/slapd

I think that is the only file with anything reporting as to zimbra's boot status.

What the heck just happened?

p.s. I've also been wondering, and have not yet found, is there any solid documentation on that what log files report what for zimbra?

Edit: If it's any help this is the output of sh -x /zimbrahome/bin/ldap start

[zimbra@lima ~]$ sh -x bin/ldap start
+ PID=
+ PIDFILE=/opt/zimbra/openldap/var/run/slapd.pid
+ mkdir -p /opt/zimbra/openldap/var/run/
+ case "$1" in
+ start
+ checkrunning
+ getpid
+ '[' -f /opt/zimbra/openldap/var/run/slapd.pid ']'
+ '[' x = x ']'
+ RUNNING=0
+ '[' 0 = 0 ']'
+ '[' '!' -f /opt/zimbra/openldap-data/DB_CONFIG ']'
+ /opt/zimbra/libexec/zmconfigure /opt/zimbra/conf/slapd.conf.in /opt/zimbra/conf/slapd.conf
+ '[' -x /opt/zimbra/sleepycat/bin/db_recover ']'
+ /opt/zimbra/sleepycat/bin/db_recover -h /opt/zimbra/openldap-data -v
db_recover: Finding last valid log LSN: file: 1 offset 5488355
db_recover: Recovery starting from [1][5488231]
db_recover: Recovery complete at Thu May 18 15:05:40 2006
db_recover: Maximum transaction ID 80000016 Recovery checkpoint [1][5488355]
+ sudo /opt/zimbra/openldap-2.3.20/libexec/slapd -l LOCAL0 -4 -u zimbra -h 'ldaps:// ldap://:389/' -f /opt/zimbra/conf/slapd.conf
+ sleep 2
+ getpid
+ '[' -f /opt/zimbra/openldap/var/run/slapd.pid ']'
+ '[' x = x ']'
+ echo 'ERROR - failed to start slapd'
ERROR - failed to start slapd
+ echo ''

+ exit 1

[jholder]

What does
>ps auxww | grep slapd

and

>ps auxww | grep zimbra


return?

ALSO: before you proceed, it would be very wise to have a backup...

[vaneck]

Originally Posted by wannabetenor

What does
>ps auxww | grep slapd

and

>ps auxww | grep zimbra


return?

ALSO: before you proceed, it would be very wise to have a backup...

They return nothing. I had at that time made sure that everything for zimbra was stopped (there were a few straggling processes), and have since posting rebooted the server twice.

And good idea about the backup. I really don't need to lose any info in doing this as well.

[jholder]

[QUOTE=vaneck] If it's any help this is the output of sh -x /zimbrahome/bin/ldap start [QUOTE]

What's the "zimbrahome" ?

[vaneck]

[QUOTE=wannabetenor][QUOTE=vaneck] If it's any help this is the output of sh -x /zimbrahome/bin/ldap start

What's the "zimbrahome" ?

it's really /opt/zimbra I gues I should've really said zimbrahomedirectory

[jholder]

Okay,
So I assume you have your backup.

I've been looking around to see what I can find about this, but haven't been able to fund much.

Unless someone has a better idea: try a install OVER your current version.

MAKE SURE YOU HAVE A SAFE BACKUP!!

But I would try to run the installer again.

Can I also get your specs?

Thanks,
jh

[vaneck]

um, apparently something I did wiped the slapd.crt file. The other guy that does NetAdmin at our company reran zmcreatecert (which I forgot was not a zmprov command) and it fixed the problem actually. I think he's going to get on here later and try to go into more detail.