Re: slapd syncrepl consumer having permanent high CPU load

[John Morrissey <jwm@horde.net>]

On Tue, Mar 10, 2009 at 09:48:20AM -0700, Howard Chu wrote:
> John Morrissey wrote:
> > After a couple days of slapd uptime, its BDB entry and DN caches have
> > grown to about 400k entries each
> 
> Grown to 400k, starting from what?

>From 0 (since slapd was started). AFAICT there is no plateau, it just keeps
growing without bound until slapd needs to be restarted since its memory
consumption has grown beyond the machine's capacity.

> What are the configured sizes, what's the total DB size?

The database is a bit over a million entries. cachesize is set to 10000
(which is probably too low). We don't explicitly set dncachesize, and the
man page says it defaults to 2*cachesize, or 20k (which is also probably too
low).

john
-- 
John Morrissey          _o            /\         ----  __o
jwm@horde.net        _-< \_          /  \       ----  <  \,
www.horde.net/    __(_)/_(_)________/    \_______(_) /_(_)__