We are having constant random YP errors on our YP clients. The errors
are always one of the following two:

yp_all: clnt_call: RPC: Timed out

YPBINDPROC_DOMAIN: Domain not bound

I get these in the output of cron jobs that run on clients and often
therefore fail. It also happens interactively but rarely. Immediately
after an error all seems to work okay when I interactively check.

We have about 200 clients (mostly Linux RedHat with a few Solaris boxes).
We have four NIS servers all Linux RedHat 7.3 with the latest RPM
updates running ypserv 2.8-0.73E.

Occasionly we have the ypserv process go crazy with messages:

WARNING(ypproc_all_2_svc): too many running children!

This seems to happen more often on the master. When this is going
on, the clients of that server are usually hung. Doing a simple
restart of ypserv seems to make everyone happy nearly instantly.

I tried increasing the cached file handles in the ypserv.conf to 60
but that seems to have had no effect on our problem.

Our networking hardware-wise appears to be fine.

Anyone have any clues as to what may be wrong?
Thanks.

--
---------------------------------------------------------------
Paul Raines email: raines@nmr.mgh.harvard.edu
MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging
149 (2301) 13th Street Charlestown, MA 02129 USA

Paul Raines <raines@nmr.mgh.harvard.edu> writes:

*snip*

Probably a good idea to throw in a few more NIS slaves.
Balance out the load manually binding if not by subnet.
NIS doesn't handle loads very well, 50 per is pushing it
I tried to keep it 25-30 clients per NIS server.


--
Vincent Fox
Georgia Institute of Technology, Atlanta Georgia, 30332
uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!vf5
Internet: vf5@prism.gatech.edu