Re: slurpd 2.1.22+db4.1.25+RH7.2-3 blocks slave' threads (too switch context x86)

[Ignacio Coupeau <icoupeau@unav.es>]

Hansjoerg Maurer wrote:

> Hi,
>
> I had some problems with signal handling and threads on newer RH kernels 
> (with smbmount, not with ldap)
> Setting
> export LD_ASSUME_KERNEL=2.2.5
> in the startup-skript (of slapd in your case) help with my problems.
> Just an idea, I don't know, if it may help in your case...

I tested:
1. the ext3 fs  with and w/o LD_ASSUME_KERNEL=2.2.5: slave ldap clogged.
2. the fs ext2: more robust, but at last, the slave-slapd is also clogged.
3. I compiled the ldbm support, re-import the db, and with a load 
generated with several scripts writing (4 scripts update 20.000 
objects), several times, I can't block the slave-slapd at all.

so at this moment, our system remains:
1. the master: "database bdb"
2. the slave(s): "database ldbm"

The only test omitted (I think) was set the slurpd with a conf-file for 
each replica.

IMHO something goes wrong with the db-4.1.25 threads or the comunication 
between linux RH <--> db-4.1.25, because at last, the db is badly corrupted.

Ignacio
--
____________________________________________________
Ignacio Coupeau, Ph.D.     icoupeau@unav.es
CTI, Director              icoupeau@alumni.unav.es
University of Navarra      icoupeau@ieee.org
Pamplona, SPAIN            http://www.unav.es/cti/