Re: commit: ldap/servers/slapd/tools slapcommon.c

["Pierangelo Masarati" <ando@sys-net.it>]

> Today at 4:03pm, Pierangelo Masarati wrote:
>
>>
>> > On Mar 3 at 2:36pm, Kurt D. Zeilenga wrote:
>> >
>> >> I suggest that no options always refer to the "same" database for
>> all (current and hopefully future) tools.   Instead of
>> >> saying the default is "-n 1", we want to say the default is
>> >> the first backend which is not subordinate and not a monitor
>> >> backend.  I guess I'm fine with that... as long as you think
>> >> no one is ever going to implement slapcat'ing (or other tools) for
>> the monitor backend.
>> >
>> > Just to play devil's advocate...
>> >
>> > Wouldn't being able to slapcat the monitor backend allow me to do
>> periodic slapcats and compare the output to see the changes so I can
>> do trend analysis/graphing?
>>
>> Slapcat is intended to read the contents of a db while the server is
>> not running.  The monitor database does not exist, it's held in a
>> cache while the server is running, and its entries that are not
>> constant (e.g. based ont he configuration at startup) are updated on
>> request.  This is called "ldapsearch".
>
> Drat.... but I can slapcat the bdb backend while the server is
> running....

If you can figure out (and implement :) an easy way for slapcat
to read slapd's memory while slapd is running, e.g.

system("ldapsearch -x -b 'cn=monitor' -LLL\n");

then I'm fine.

p.

-- 
Pierangelo Masarati
mailto:pierangelo.masarati@sys-net.it