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

slapd crashing when setting cachesize too high (ITS#1265)



Full_Name: Markus Storm
Version: REL_ENG_2 as of 07/31/01
OS: Linux 2.4
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (193.189.224.37)


slapd crashes on searches for "all entries" when I set cachesize too high:

2001-07-31 16:25:27.306008500 ber_flush: 49 bytes to sd 20
2001-07-31 16:25:27.306642500 ch_calloc of 17488 elems of 4 bytes failed
2001-07-31 16:25:27.306709500 slapd: ch_malloc.c:77: ch_calloc: Assertion `0'
failed.
2001-07-31 16:25:27.306843500 ber_flush: 49 bytes to sd 20

For our database, the limit seems to be somewhere between 115.000 and 120.000.


We have a ~ 400.000 user database with an average entry size of about 700
bytes.
Among others, this happens on a Linux 2.4 machine with 4 GB RAM. 
OS limits (heapsize, swapsize, process size, etc etc.) are NOT hit.

This is reproducible. slapcat/slapadd'ing the whole database doesn't make any
difference.