Followup: Cluster 5.1-problem

From: Danny Petterson <danny.petterson_at_framfab.dk>
Date: Thu, 29 Mar 2001 20:49:10 +0200

....a weird thing is, that if I kill node 1 Oracle does not fail over, it
just figures as if nothing has happend (ONLINE on the failed node, if you
look at caa_stat). When node 1 then is booted again, Oracle starts up on
this, and everything resumes in a normal fashion. Its even possible to
relocate to node 2 (so nothing in the cluster has a clue, that node 1 has
been down).


Med venlig hilsen/Kind Regards
         
Danny Petterson Driftbolaget A/S.
System Engineer HTTP://WWW.DRIFTBOLAGET.COM
TLF: +45 21716422 Danny.Petterson_at_Driftbolaget.com
                        
 
        "An expert is a man who has made all the mistakes
         which can be made in a very narrow field"
                                                Niels Bohr





-----Oprindelig meddelelse-----
Fra: Jason Orendorf [mailto:orendorf_at_alf.dec.com]
Sendt: Thursday, March 29, 2001 8:12 PM
Til: Danny Petterson
Emne: Re: Cluster 5.1-problem


Danny Petterson wrote:
> When running
> #clu_check_config
> I get these ATTENTION-lines:
...
> ***** Checking daemons on members *****
> *****
> ***** Checking member tranodcph102-mc0 daemons
> aliasd is NOT_RUNNING !!!!!ATTENTION!!!!!
...

A few things to check:

- make sure /.rhosts contains all cluster member names so
  rsh works between members (try entering all hostnames, all
  MC interface names, and the cluster alias)
- make sure /etc/inetd.conf has the line for "cfgmgr"
  uncommented (also avoid using a TCP wrapper for this service)

Hope this helps.

 Jason Orendorf - Compaq Tru64 Unix Support
 Customer Support Center
 orendorf_at_alf.dec.com
 1(800)354-9000
Received on Thu Mar 29 2001 - 18:50:28 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:42 NZDT