NIS+ problems after network drop

NIS+ problems after network drop

Post by def » Thu, 26 Aug 1999 04:00:00



I am using NIS+ with the NIS+ server and client on different subnets;
both Solaris 2.6; for now I do not have a replica.  At some point in the
night, network connectivity was lost to the sight that has the client.
After networking was re-established, I was still not able to log in to
the client (telnet and rsh commands would just hang).  While usually I
would not be able to, I thought the easy way out would be to reboot both
the server and the client; server rebooted fine but the client hung
during boot after the NIS+ daemons were starting.

On the server, the only messages in /var/adm/messages are:
Aug 25 05:22:30 cmosun30-1 nisd[156]: Object
hosts.org_dir.cmo.census.gov. isn't a directory.
Aug 25 05:22:32 cmosun30-1 last message repeated 123 times
NOTE: this message STILL comes out when I do a "niscat hosts"; even
though the command works - any ideas?

On the client, there are many messages, summarized:
Aug 25 03:00:53 cmosun5-2 nis_cachemgr: NIS+ server not responding
Aug 25 03:02:31 cmosun5-2 last message repeated 2 times
Aug 25 03:49:12 cmosun5-2 cron[23447]: authdes_seccreate: keyserv(1m) is
unable to generate session key

These messages stopped happening after the network came back but I could
not connect to the client machine; I was able to ping it.

Questions:

1. If I had a replica on the same subnet as the client - would it/should
it have taken over for the server and all would have been happy?

2. Is there a need to reset the credentials?  I found some info in a FAQ
(title: How to fix NIS+ server breaks with Unable to Authenticate NIS+
server) that had a set of commands including stopping and starting the
nisd and keyserv daemons, and running nisupdkeys, etc. - seems like a
lot to go through just because the network dropped!

I was trying several commands on the NIS+ server (not the ones mentioned
in #2, but I had already rebooted twice so part of that was done), and
the client finished booting and all was well.  But now I am not sure
which command did it!!??  I ran
"nisclient -c -d cmo.census.gov cmosun5-2" on the server to try to reset
the client; it said "skipped, already existed" but I wonder if it still
reset something.  Other than that, I don't see any commands that I ran
that would have helped except the reboots.

Any help would be appreciated.

--
*******************************

Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

 
 
 

1. NIS+ problems after network drop

I am using NIS+ with the NIS+ server and client on different subnets;
both Solaris 2.6; for now I do not have a replica.  At some point in the
night, network connectivity was lost to the sight that has the client.
After networking was re-established, I was still not able to log in to
the client (telnet and rsh commands would just hang).  While usually I
would not be able to, I thought the easy way out would be to reboot both
the server and the client; server rebooted fine but the client hung
during boot after the NIS+ daemons were starting.

On the server, the only messages in /var/adm/messages are:
Aug 25 05:22:30 cmosun30-1 nisd[156]: Object
hosts.org_dir.cmo.census.gov. isn't a directory.
Aug 25 05:22:32 cmosun30-1 last message repeated 123 times
NOTE: this message STILL comes out when I do a "niscat hosts"; even
though the command works - any ideas?

On the client, there are many messages, summarized:
Aug 25 03:00:53 cmosun5-2 nis_cachemgr: NIS+ server not responding
Aug 25 03:02:31 cmosun5-2 last message repeated 2 times
Aug 25 03:49:12 cmosun5-2 cron[23447]: authdes_seccreate: keyserv(1m) is
unable to generate session key

These messages stopped happening after the network came back but I could
not connect to the client machine; I was able to ping it.

Questions:

1. If I had a replica on the same subnet as the client - would it/should
it have taken over for the server and all would have been happy?

2. Is there a need to reset the credentials?  I found some info in a FAQ
(title: How to fix NIS+ server breaks with Unable to Authenticate NIS+
server) that had a set of commands including stopping and starting the
nisd and keyserv daemons, and running nisupdkeys, etc. - seems like a
lot to go through just because the network dropped!

I was trying several commands on the NIS+ server (not the ones mentioned
in #2, but I had already rebooted twice so part of that was done), and
the client finished booting and all was well.  But now I am not sure
which command did it!!??  I ran
"nisclient -c -d cmo.census.gov cmosun5-2" on the server to try to reset
the client; it said "skipped, already existed" but I wonder if it still
reset something.  Other than that, I don't see any commands that I ran
that would have helped except the reboots.

Any help would be appreciated.

--
*******************************

Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

2. routed + Novell RIP

3. Informix problems afters Y2K Solaris 2.5 patch

4. connecting to the internet

5. ipsec problem: shunt SA of DROP or no eroute: dropping.

6. Web server hardware architecture?

7. NIS : auth problem with Linux nis server and SUN sparc nis client

8. ignore this also

9. Problem with computer network dropping connection

10. callback, ppp, and morning-afters

11. Question: How to use NIS+ or NIS in a alread established Network

12. NIS and NIS+ Simultaneously on Same Network?

13. NIS+ vs NIS for a small network