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

slapadd bulk load performance issue



Hello,

 

Version OpenLDAP: 2.4.11

OS: Linux Red Hat ES 4.6

Storage: Berkeley DB 4.2 (+4 patches)

 

 

I'm testing performance in importing big LDIF file with 2 milions of entries (22.000.000 number of attributes). LDIF file is 600 MB big, one entry looks like this:

 

dn: itContainerId=20000001,ou=Container,l=example,c=com

objectClass: labeledURIObject

objectClass: itContainerOC

itContainerId: 20000001

itContainerType: 1

itContainerName: ljubljana1

itContainerStatus: 0

itParentContainerId: 3000000

itConfigurationNeeded: FALSE

itSerialNumber: 04

itRegType: 1

 

 

I use slapadd tool with -q option :

 

slapadd -q -f slapd.conf –l 2_milions_of_entry.ldif

 

and the following configuration settings:

 

database bdb

dbcachesize 1000000

cachesize 150000

sizelimit 150000

checkpoint      10    1

 

set_cachesize 0 20000000 0

set_lg_regionmax 262144

set_lg_bsize 2097152

 

 

In my case, it takes about 20!!! hours to load the entire LDIF contents. Is this normal? Did anybody has experiences with loading of such amount of data?

Am i perhaps having wrong settings in config file? Can you please someone tell me, how much time is normaly needed for slapadd?

 

Thanks for your help.

 

Best Regards

Blaz Skulj