IBM SG24-5131-00 Manuale Utente

Cluster Testing 

  • Check that all interfaces communicate (

ping <ip-address> 

or 

ping -R 

<ip-address>).

  • List the arp table entries with 

arp -a

.

  • Check the status of the TCP/IP daemons (

lssrc -g tcpip

).

  • Ensure that there are no bad entries in the /etc/hosts file, especially at the 

bottom of the file.

  • Verify that, if DNS is in use, the DNS servers are correctly defined (

more 

/etc/resolv.conf)

.

  • Check the status of NIS by typing 

ps -ef | grep ypbind

 and 

lssrc -g yp.

  • The command 

exportfs

 shows non-HACMP controlled NFS exports.

  • Run 

snmpinfo -m dump -o /usr/sbin/cluster/hacmp.defs address 

show 

snmp information for Cluster network addresses (including the serial 
interfaces).

  • Ensure that the correct VG's are defined, that quorum and auto-varyon are 

correctly defined, and that the shared VG's are in the correct state (

lsvg

 

and 

lsvg -o

).

  • Check that there are no stale partitions (

lsvg -l)

.

  • Check that all appropriate file systems have been mounted and that none 

of the rootvg file systems are full (

df -k

).

  • Check that PVid's have been assigned where necessary and that there 

are no ghost disks(

lspv

).

  • Verify that all entries in the /etc/filesystems file are correct and that there 

are no erroneous entries (

more /etc/filesystems

 and 

lsfs

).

  • Check the status of the cluster daemons by issuing 

lssrc -g cluster

 and 

lssrc -g lock

.

  • Run 

/usr/sbin/cluster/clstat

 to check the status of the cluster and the 

status of the network interfaces.

  • Check the cluster logfiles with 

tail -f /tmp/hacmp.out

, 

more 

/usr/sbin/cluster/history/cluster.mmdd

 (mmdd = current date), 

tail -f 

/var/adm/cluster.log

 and 

more /tmp/cm.log

.

  • Check that the nodename is correct (

odmget HACMPcluster)

.