This is the mail archive of the
mailing list for the glibc project.
Re: NIS performance patches
- From: Jeff Bastian <jmbastia at ti dot com>
- To: Thorsten Kukuk <kukuk at suse dot de>
- Cc: libc-alpha at sources dot redhat dot com
- Date: Thu, 12 Feb 2004 21:18:07 -0600
- Subject: Re: NIS performance patches
- References: <3FFF4DE9.email@example.com> <20040115175855.GA9152@suse.de> <20040118205618.GA25908@suse.de> <20040203152140.GA4971@suse.de> <40227BB9.firstname.lastname@example.org> <402C2926.email@example.com>
Jeff Bastian wrote:
I finally managed to build a NIS server I can play with. I fired up
'ypbind -no-ping' on my client and it bound to the new server. Then I
killed ypserv. The client broke:
Ignore my previous e-mail. The client works fine. It was a long day at
work and I put a bad IP address in my yp.conf file...
So, now I have a yp.conf file with two servers: my test box and a
production NIS server. I wrote a script that'll keep restarting ypbind
until it picks my test server. Then I run ypwhich, ypmatch, and ypcat
to verify it works.
I kill the ypserv daemon on the test server. I run ypwhich and it
immediately rebinds to the production server.
I repeat the process, except when I kill ypserv, I try ypmatch on the
client. This spits out an error message, but then rebinds to the
production server and finds the record I'm trying to match.
I repeat the process again, except after killing ypserv, I try ypcat.
This fails and does not cause the client to rebind.
This is all using glibc-2.3.2 and ypbind-1.12 as shipped with RedHat
Enterprise Linux 3 with Update 1 plus the NIS patches.
I haven't looked at why ypcat fails to trigger a rebind yet. Hopefully
I can find some time tomorrow afternoon to do that.