[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: slapd crashes (ldap 2.1.30, db 4.1.25)

Hash: SHA1

John Borwick wrote:
| Hello.  We just moved our production LDAP server to OpenLDAP, after lots
| of good results running in limited tests.
| *Something* is crashing slapd.  I don't know what.  Nothing's logged,
| and core isn't dumped.  I will submit an ITS if I can figure out what's
| going on.
| Our setup:
| ~  Red Hat Workstation 3.0 w/ 2 CPUS ( + SMP ) and 2 Gb RAM
| ~  o=WFU,c=US
| ~    is an LDAP backend
| ~    "suffixmassage" translating to "ou=Users,dc=wfu,dc=edu"
| ~  dc=wfu,dc=edu
| ~    is a BDB backend.
| ~    No DB_CONFIG, "cachesize" set to 5000 in slapd.conf
| Our secondary server is the one everyone's hitting.  It seems like a lot
| of connections aren't being closed by the clients; maybe that has
| something to do with it.
| I've turned logging up A LOT but there's no consistent error.  Some of
| the things that show up in the log include
| connection_close: deferring
| ber_get_next on fd 172 failed errno=11 (Resource temporarily unavailable)
| My questions for you:
| ~  * does anyone know why slapd is crashing?
| ~  * is there anything I *need* in DB_CONFIG or slapd.conf, e.g. an idle
| timeout?
| ~  * are there any configuration parameters that, when improperly
| configured, can cause slapd to crash?
| ~  * are there any file permissions, etc. on the UNIX side that could
| cause slapd to crash?

Things to try:
- -Berkeley db>=
- -ensure you have a checkpoint setting in your database definition in

If you plan to never restore slave LDAP db's from backups, turn off
transaction loggin on the slave via the DB_CONFIG (and instead, ensure
you are treating your transaction logs on the master very well).


- --
Buchan Milne                      Senior Support Technician
Obsidian Systems                  http://www.obsidian.co.za
B.Eng                                RHCE (803004789010797)
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org