SUMMARY: NIS Update Problem

From: John M. Vogtle (jmvogtle@cadem.mc.xerox.com)
Date: Tue Nov 21 1995 - 00:30:18 CST


Greetings,

I am now able to update my NIS maps to my remote sites. Special
thanks to the following for their input:

John Stewart <jns@cisco.com>
Anatoly.Lisovsky@kamaz.kazan.su (Anatoly M. Lisovsky)
celeste@stokely.com (Celeste Stokely)
manjeet@iglou.com (Manjeet Rekhi)
stern@sunrise.east.sun.com (Hal Stern - Distinguished Systems Engineer)

I probably wasn't real clear in that the netresources map had been
working fine until relatively recently. It is a home brewed map that
we use for a variety of functions.

SOLUTION:

John Stewart <jns@cisco.com> hit the nail right on the head:

    Ace> We've seen this behaviour over low speed links and there was
    Ace> line protocol failure because of a framing error.

Our local telco swapped out a bad mux and things are peachy.

My original message:

    Original> Greetings,

    Original> I've run into a rather bizarre NIS update problem that I
    Original> can't seem to diagnose. I appeal to the collective
    Original> wisdom of the net to come to my aid. (PLEASE)

    Original> I'm trying to update a NIS map to a remote site over a
    Original> T1 line. This particular map is non-standard but works
    Original> fine in all our other sites. On the server I get the
    Original> following message:

    Original> wizard# make netresources updated netresources

    Original> Status received from ypxfr on morgan: Failed - ypxfr
    Original> had an RPC failure Status received from ypxfr on tyler:
    Original> Failed - ypxfr had an RPC failure pushed netresources

    Original> morgan and tyler are the servers at the other end of the
    Original> T1. Note that there are local servers that recieve the
    Original> update just fine. Also (and most bizarre), the
    Original> netresources map is the ONLY NIS map that morgan and
    Original> tyler have trouble recieving - maps like hosts,
    Original> netgroup, etc. work just fine.

    Original> If I try to ypxfr the map on either machine, I get
    Original> messages like the following:

    Original> morgan# /usr/etc/yp/ypxfr -f -h wizard netresources
    Original> clnt call to ypxfrd getdbm failed. getdbm: RPC: Timed
    Original> out (info) wizard cadem.mc.xerox.com netresources ypxfrd
    Original> getdbm failed (reason = -3) -- using ypxfr

    Original> Again, something like "ypxfr -f -h wizard hosts.byname"
    Original> works just fine.

    Original> All workstations in question are running SunOS 4.1.3_U1.

Other suggestions are included below. Thanks to one and all

>From Anatoly M. Lisovsky:

    Anatoly> Hi John, i think Secure RPC wo'nt work because of non
    Anatoly> stable time delays.

From: celeste@stokely.com (Celeste Stokely)

    Celeste> I may be ignorant, but I've never heard of an NIS map of
    Celeste> "netresources" and neither has my 4.1.4 NIS
    Celeste> master. Perhaps you've added this into your master's maps
    Celeste> and haven't updated the NIS Makefile to support it on
    Celeste> morgan and tyler?

From: manjeet@iglou.com (Manjeet Rekhi)

    Manjeet> You may want to delete these maps from secondary
    Manjeet> servers(morgan and tyler) and retry.

From: stern@sunrise.east.sun.com (Hal Stern - Distinguished Systems Engineer)

    Hal> do you have the netresources map already on the slave
    Hal> servers? that is, does some copy of it exist there? if not,
    Hal> the ypxfr will fail when it attempts to send a "clear all"
    Hal> message to the slave server -- it will not have any map to
    Hal> clear the entries from. you'll probably need to explicitly
    Hal> drag the map across (specifying the master name, the domain
    Hal> name, etc) and then ypxfr should be happier.

-- 

John M. Vogtle Internet: jmvogtle@cadem.mc.xerox.com UNIX Systems Programmer Phone: (716) 422-6475 US Mail: 800 Phillips Rd. Bldg 147-59c "One planet is all you get." Webster, New York 14580



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:35 CDT