unexpected slapd shutdown

[maumar]

This is

Release 7.1.0_GA_3140.RHEL5_20110329150824 CentOS5 FOSS edition.

It was zimbra 5 upgraded to 6 and then to 7, running smoothly by 2 years, without ever a single issue.

This morning at 9:45 o'clock it decided to do karakiri, sadly;

May 10 09:45:49 zimbra slapd[4396]: daemon: shutdown requested and initiated. 
May 10 09:45:57 zimbra slapd[4396]: slapd shutdown: waiting for 0 operations/tasks to finish 
May 10 09:45:59 zimbra slapd[4396]: slapd stopped. 
May 10 09:46:07 zimbra postfix/smtpd[1437]: connect from localhost.localdomain[127.0.0.1]
May 10 09:46:07 zimbra postfix/proxymap[30845]: error: dict_ldap_connect: Unable to set STARTTLS: -1: Can't contact LDAP server

nothing more, nothing less.
The Geotrust ssl cert was installed one week ago, without any issue.
IMHO i need a more verbose logging level to do a diagnose.
Which is the right level of log to diagnose this kind of problem?
Is 256 like wiki suggest

http://wiki.zimbra.com/index.php?title=Change_OpenLDAP_slapd_Logging

the proper value?
What do you suggest to set?

[resol]

Hi maumar,

I have exactly the same problem, happened now already 3 times during the last month. Have you found out what might be the problem here already?

After I stop and start the services everything is fine again until this happens again.

I have the latest zimbra version running on ubuntu LTS.

Thanks a lot!
Stephan

[maumar]

No, i have not solved.
I have to increase ldap logging verbosity, but until now I was too busy.
With default verbosity there is no chance to diagnose anything useful

[resol]

Anyway thanks for your reply!

For me it maybe seems like sldapd got the message from somebody to shutdown, so if it didn't crash will there still be something to see if you increase the loglevel?

Please keep me updated if you find out something.

Thanks a lot,
Stephan

[cvidal]

Originally Posted by maumar

May 10 09:45:49 zimbra slapd[4396]: daemon: shutdown requested and initiated. 
May 10 09:45:57 zimbra slapd[4396]: slapd shutdown: waiting for 0 operations/tasks to finish 
May 10 09:45:59 zimbra slapd[4396]: slapd stopped. 
May 10 09:46:07 zimbra postfix/smtpd[1437]: connect from localhost.localdomain[127.0.0.1]
May 10 09:46:07 zimbra postfix/proxymap[30845]: error: dict_ldap_connect: Unable to set STARTTLS: -1: Can't contact LDAP server

The same here, version 7.1.1, happened yesterday at 19:26 after more than a month of flawless uptime.

[phoenix]

Originally Posted by cvidal

The same here, version 7.1.1, happened yesterday at 19:26 after more than a month of flawless uptime.

Have you checked if you're certificates have expired? You could also check some of the solutions in these threads: site:zimbra.com +"dict_ldap_connect: Unable to set STARTTLS" - Yahoo! Search Results

[nedm]

We're on 7.1.2_GA_3268/Ubuntu 10.04.3 and I see the same thing in zimbra.log before everything goes haywire. A simple

zmcontrol start

starts up the affected services and everything runs smoothly, but the random nature and timing of the problem and the inability to see any root cause after combing the logs in /var/log and opt/zimbra/log makes me worried.

Anyone else with any input would be greatly appreciated.

[vinzenz]

I have exactly the same problem: sldapd just stops and the whole mail system follows soon after.
I had increased log verbosity of ldap to 512, but still all it says is

Dec  8 05:38:05 www slapd[22100]: daemon: shutdown requested and initiated.
Dec  8 05:38:06 www slapd[22100]: slapd shutdown: waiting for 0 operations/tasks to finish
Dec  8 05:38:07 www slapd[22100]: slapd stopped.

I have no idea who initiated the shutdown and it seem to be unpredictable.
Is this issue being investigated? I'd be happy to help if I can.

[sinmrki]

The same problem here, have anyone some solution or idea please? thank you.

[fseaton]

Same here, too. Zimbra master ldap server has been running flawlessly since last upgrade (in January to 7.1.3_GA_3346) and at 8 am this morning it just shut itself down:

Mar 21 08:04:09 zcs1 slapd[19983]: daemon: shutdown requested and initiated.
Mar 21 08:04:09 zcs1 slapd[19983]: slapd shutdown: waiting for 0 operations/tasks to finish
Mar 21 08:04:11 zcs1 slapd[19983]: slapd stopped.


Any ideas out there?