Mutex problems in solaris using openldap

[Chuck Pierce <chuck@onyxsys.net>]

Hey, has anyone else been experiencing this problems under Solaris8 and experiencing
2.0.21??

I've got openldap 2.0.21 under solaris 8 (kernel patch 17) and the box is
getting creamed...

I don't have that many records (under 2000), but it is getting hit pretty heavily
by 8 web servers.  The box is doing nfs and ldap and is getting a 0% idle time
about 90% of the time (20% is user and 80% is kernel).

After doing some research (and mxstat) I found that I was doing ~1500 mutex
waits per second per processor (it's a dual processor system).

Any help/advice would be greatly appreciated.

thanks - Chuck