Re: RE24: Locker killed to resolve a deadlock

[Howard Chu <hyc@symas.com>]

Michael Ströder wrote:
> HI!
>
> Playing with recent RE24 and MIT Kerberos' LDAP backend I noticed this
> in syslog:
>
> Oct 11 10:15:35 nb2 slapd[3888]: =>  bdb_dn2id_add 0x8d: put failed:
> DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock -30995
> Oct 11 10:15:35 nb2 slapd[3888]: =>  bdb_dn2id_add 0x8c: put failed:
> DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock -30995
> Oct 11 10:15:35 nb2 slapd[3888]: conn=5 fd=29 closed
> Oct 11 10:15:35 nb2 slapd[3888]: =>  bdb_dn2id_add 0x8d: put failed:
> DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock -30995
>
> Can't reproduce though. Under which circumstances can this happen?

Various multiple access scenarios. It's not an error, that's normal in a 
transactional environment. The operations are automatically retried, ignore 
these messages.

--
  -- Howard Chu
  CTO, Symas Corp.           http://www.symas.com
  Director, Highland Sun     http://highlandsun.com/hyc/
  Chief Architect, OpenLDAP  http://www.openldap.org/project/