[Kolab-devel] [issue1132] kolabd/slurpd crash every day under high numer of users

[Fabio Pietrosanti / Khamsa SA]

New submission from Fabio Pietrosanti / Khamsa SA <kolab at khamsa.ch>:

Under high number of users (78k) kolabd and slurpd seems to crash without giving usefull 
informations.

Here is the kolabd debugging log of the master server used only for ldap and administration:

user:Feb 14 15:56:55 ldap1-be-rm kolabd[23016]: L: Finished synchronisation
user:Feb 14 15:56:55 ldap1-be-rm kolabd[23016]: K: Synchronisation complete, starting up daemon
user:Feb 14 15:56:55 ldap1-be-rm kolabd[23016]: K: Forking `dirservd' listener
user:Feb 14 15:56:55 ldap1-be-rm kolabd[23016]: K: Forking `slurpd' listener
user:Feb 14 15:56:55 ldap1-be-rm kolabd[23016]: K: Listeners spawned, wait()ing
user:Feb 14 15:56:55 ldap1-be-rm kolabd[25586]: SD: Opening listen server on 127.0.0.1:9999
user:Feb 14 15:56:55 ldap1-be-rm kolabd[25586]: SD: Listen server opened, waiting for incoming 
connections
user:Feb 14 15:56:55 ldap1-be-rm kolabd[25585]: DSd: Listener starting up, refresh is: 120 seconds
user:Feb 14 15:56:55 ldap1-be-rm kolabd[25585]: K: `dirservd' returned
user:Feb 14 15:56:57 ldap1-be-rm kolabd[23016]: K: Child terminated normally
user:Feb 14 15:59:44 ldap1-be-rm kolabd[25586]: SD: Incoming connection accepted, peer=127.0.0.1
user:Feb 14 15:59:44 ldap1-be-rm kolabd[25586]: SD: Waiting for LDAP updates
user:Feb 14 15:59:44 ldap1-be-rm kolabd[25586]: SD Debug: Reading ASN
user:Feb 14 15:59:45 ldap1-be-rm kolabd[25586]: SD Debug: Reading ASN
user:Feb 14 15:59:45 ldap1-be-rm kolabd[25586]: SD Debug: Connection closed
user:Feb 14 15:59:46 ldap1-be-rm kolabd[25586]: SD Debug: Request bindRequest received
user:Feb 14 15:59:46 ldap1-be-rm kolabd[25586]: SD Debug: Writing response
user:Feb 14 15:59:46 ldap1-be-rm kolabd[25586]: B Error: Error in function 
`Kolab::LDAP::Backend::slurpd::run': Can't use an undefined value as a symbol reference at /kolab/lib/
perl/site_perl/5.8.7/i686-linux/Kolab/LDAP/Backend/slurpd.pm line 516. , exiting
user:Feb 14 15:59:47 ldap1-be-rm kolabd[23016]: K: Abnormal child exit status encountered, aborting
user:Feb 14 15:59:47 ldap1-be-rm kolabd[23016]: Kolab is shutting down
user:Feb 14 15:59:47 ldap1-be-rm kolabd[23016]: K: SIGINT/SIGTERM detected, kill()ing children
user:Feb 14 15:59:47 ldap1-be-rm kolabd[23016]: L: Shutting down
user:Feb 14 15:59:47 ldap1-be-rm kolabd[23016]: K: Exiting

I found that kolabd and slurpd are died and i'm wondering if there is a strict dependancy between this 
two components.
Could be possible that kolabd die and then slurpd die or the reverse?

How can i enable more debugging in order to find out the problem?

Kolab installation use the latest kolabd code available for 2.1 branch.

----------
messages: 6576
nosy: khamsa
priority: critical
status: unread
title: kolabd/slurpd crash every day under high numer of users
________________________________________________
Kolab issue tracker <kolab-issues at intevation.de>
<https://intevation.de/roundup/kolab/issue1132>
________________________________________________