Hello,
Thank you for all the responses. I thought I'd summarize now, even though I have
not yet seen my original query or the apology for the bouncing of mail in my
incoming mail box.
The problem was that diskless clients would hang after the server had been
rebooted until a successful login had been achieved, or the client had also been
rebooted. Many people wrote saying that they have the same problem.
Apparently the problem is due to a bug in the ypbind program and its caching of
the ypserver information and its very long timeout for this cached information.
There is an official SUN patch avaiable which is basically a new ypbind program.
The patch number is 100342-01 and here is the readme:
Date: 1/Aug/1991
SunOS release: 4.1 4.1.1
Unbundled Product:
Unbundled Release:
Topic: NIS rebind patch
BugID'd fixed for this patch: 1046416
Architectures for which this patch is available: sun3, sun4
Patches which may conflict with this patch:
Obsoleted by: Sys_V_Rel4
Problem Description:
Bug 1046416:
If you bring a ypserver down into single user and then boot it into
multi user by either typing control D or reboot, the yp clients
will take a long time to rebind to the server.
********************* WARNING ******************************
This is a new version of ypbind that never uses the NIS
binding file to cache the servers binding. This will have
the effect of fixing the current symptom. However, it might
degrade the overall performance of the system when the
server is available. This is most likely to happen on an
overloaded server, which will cause the network to produce
a broadcast storm.
*************************************************************
INSTALL::
As root and for the correct architecture directory.
#kill the currently running ypbind
kill <processid of ypbind>
mv /usr/etc/ypbind /usr/etc/ypbind.FCS
#copy the new version to /usr/etc
cp sun{3,4}/ypbind /usr/etc
chown root /usr/etc/ypbind
chmod 755 /usr/etc/ypbind
restart ypbind
I haven't installed this yet, since our network is quite busy and I wasn't
very happy about the warning.
The responses were as follows:
Me Too:
d.g.checketts@uk.ac.bham
terry@esrg.ucsb.edu
debbie@cs.odu.edu
matt%com.xerox.wbst845e@alpha.xerox.com
lamour@mitre.org
john@rosemount.co.uk
brian@mrc-bsu.cam.ac.uk
dal@gcm.com
jpd@usl.edu
kim@cs.clemson.edu
The Patch:
miker@sbcoc.com
mike@trdlnk.uucp
as6143@eerpf001.ca.boeing.com
D.Mitchell@dcs.sheffield.ac.uk
Other Workarounds:
das@ee.ed.ac.uk sent a script to kill and restart ypbind on the clients if a
lock file written by the NIS server had changed (due to a reboot). If anyone
wants this contact me or him.
jp107@damtp.cambridge.ac.uk suggested manually restarting the ypbind process
on each client.
dal@gcm.com suggested using ypwhich <hung client> as the kick to get ypbind
to rebind.
Just Live With It:
mark%dm%deltam@uunet.uu.net (well, try a login, wait two minutes and all will
OK)
Other Suggested Causes:
rjwhite@watmath.uwaterloo.ca suggested that another part of the delay could be
due to /usr/ucb/quota. There has been quite a bit of comment lately about quota
causing login delays since it touches all mounted NFS filesystems and if some
NFS servers are away, login cannot complete. He suggests linking /usr/ucb/quota
to /bin/true if quotas are not needed. This will speed up normal logins as well.
If anyone wants further details of any of the mail I received, please mail.
If you want the patch, if you have SUN support, they can email the patch
to you. That is where I got it. I have not seen it on any archives. If you do
not have support, I will mail the patch to you.
Tom.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:34 CDT