HW/OS2012. 3. 20. 10:55
HW/OS2012. 3. 14. 13:41
– model and machinfo commands:
# model
ia64 hp server rx8800
# machinfo
CPU info:
Memory: 81610 MB
Firmware info:
Platform info:
OS info:
- disk information:
# ioscan -knfC disk
- lan cards information:
# ioscan -C lan
- lun information:
# ioscan -m lun
- EFI boot disk path:
# ioscan -eN | grep -i efi
- hardware health – ie FC cards:
# ioscan -P health -C fc
- full hardware listing:
# ioscan -f
- slot olrad – speed, occupancy, configuration:
# olrad -q
- lun wwid – scsimgr :
# scsimgr get_attr -a wwid -H 64000/0xfa00/0×2
# scsimgr get_attr -a wwid all_lun
– print_manifest :
# print_manifest | more
- Support Tools Manager
# cstm
> map
> sel dev 2
> info
> infolog
- Storage WWN
# ioscan -kfnC fc OR ioscan -fnC fc
# fcmsutil /dev/fcd0
- CPU
# ioscan -fknC processor | grep processor | wc -l
- CPU speed
# echo itick_per_user/D | adb -k /stand/vmunix /dev/mem | tail -1 | awk '{print $2}'
- memory
# value=`dmesg | grep Physical | grep Kbytes | awk '{print $2}'`
# expr $value / 1024
- disk
# ioscan -fknC disk | grep disk | grep -v "CD-ROM" | grep -v "DVD-ROM" | wc -l
# hwpath=`ioscan -fknC disk | grep disk | grep -v "CD-ROM" | grep -v "DVD-ROM" | head -1`
# name=`ioscan -fkn -H $hwpath | grep "/rdsk/" | awk '{print $2}'`
# value=`diskinfo $name | awk '/size:/ {print $2}'`
# expr $value / 1024
Cluster/HA2012. 2. 22. 13:53
cmapplyconf
-C
${config}
Applies the configuration files checked above.
Appears to repeat the configuration file checks, so checking them first may
be redundant.
-P ${config}
-p ${file}
cmcheckconf -C ${config} Configuration file checker. The -C checks cluster configuration files; -Pchecks package configuration files, and the -p checks all the configuration files listed in ${file} -P ${config} -p ${file}
cmdeleteconf -p Removes cluster configuration information. Use the -p option to remove package definitions from the existing cluster. I imagine there's more options; however, that one would seem to be the most used.
cmgetconf -c ${cluster} Generates configuration files that generated the existing cluster or package. Useful for identifying current configurations and for replacing configuration files that have been wiped out. Probably a good idea to use this to generate a config file to edit vs trying to keep track of changes on all cluster nodes. -p ${pkg}
cmhaltcl [ -f ] Halts a cluster. If any pakcages are running and the -f option is not supplied, the command will fail and complain about the running packages. Using the -f option will force the packages to halt before killing the cluster
cmhaltnode -f Halts a running node. Using the -f flag will switch any packages running on that particular node to an adoptive node. This command does not disable the global switch for any affected packages.
cmmakepkg -p ${config} Generates package configuration an run/halt scripts. The files generated must be edited prior to use. -p = package configuration while -s = run/halt script. -s ${script}
cmquerycl -n ${node} Use this command to set up the initial cluster configuration. Add a -noption for each node in the nascent cluster. The -C option specifies the cluster configuration file and the -v option requests verbose output. -C ${config} -v
cmruncl Starts up the cluster software and any configured packages on their primary nodes. cmrunnode Activate a previously halted node. DO NOT use this command to start up a cluster as the cluster reform process will hang. See "Hung cluster reformation" for solution to this problem.
cmscancl -s More detailed display of cluster configuration and status. -o will output to a file; -n will display information on ${node} only, and -s will display to the screen. -o ${file} -n ${node}
cmviewcl -v, Primary method for checking on the health and status of a running cluster. With no command line arguments, the command will display the cluster status (up/down) and the nodes that are attached and their status. -v = verbose with fairly obvious results. Adding other modifiers, such as -p will limit the display to the specific subsystem requested (-p = packages>. use cmviewcl -vp ${pkg} to ID global and local switch settings. -p, -c (others?)
cmviewconf -o ${file} Displays cluster configuration information. If supplied with the -o ${file}option, will output the results to ${file}. It won't overwrite ${file} if it exists.
cmcheckconf -C ${config} Configuration file checker. The -C checks cluster configuration files; -Pchecks package configuration files, and the -p checks all the configuration files listed in ${file} -P ${config} -p ${file}
cmdeleteconf -p Removes cluster configuration information. Use the -p option to remove package definitions from the existing cluster. I imagine there's more options; however, that one would seem to be the most used.
cmgetconf -c ${cluster} Generates configuration files that generated the existing cluster or package. Useful for identifying current configurations and for replacing configuration files that have been wiped out. Probably a good idea to use this to generate a config file to edit vs trying to keep track of changes on all cluster nodes. -p ${pkg}
cmhaltcl [ -f ] Halts a cluster. If any pakcages are running and the -f option is not supplied, the command will fail and complain about the running packages. Using the -f option will force the packages to halt before killing the cluster
cmhaltnode -f Halts a running node. Using the -f flag will switch any packages running on that particular node to an adoptive node. This command does not disable the global switch for any affected packages.
cmmakepkg -p ${config} Generates package configuration an run/halt scripts. The files generated must be edited prior to use. -p = package configuration while -s = run/halt script. -s ${script}
cmquerycl -n ${node} Use this command to set up the initial cluster configuration. Add a -noption for each node in the nascent cluster. The -C option specifies the cluster configuration file and the -v option requests verbose output. -C ${config} -v
cmruncl Starts up the cluster software and any configured packages on their primary nodes. cmrunnode Activate a previously halted node. DO NOT use this command to start up a cluster as the cluster reform process will hang. See "Hung cluster reformation" for solution to this problem.
cmscancl -s More detailed display of cluster configuration and status. -o will output to a file; -n will display information on ${node} only, and -s will display to the screen. -o ${file} -n ${node}
cmviewcl -v, Primary method for checking on the health and status of a running cluster. With no command line arguments, the command will display the cluster status (up/down) and the nodes that are attached and their status. -v = verbose with fairly obvious results. Adding other modifiers, such as -p will limit the display to the specific subsystem requested (-p = packages>. use cmviewcl -vp ${pkg} to ID global and local switch settings. -p, -c (others?)
cmviewconf -o ${file} Displays cluster configuration information. If supplied with the -o ${file}option, will output the results to ${file}. It won't overwrite ${file} if it exists.
HW/OS2012. 2. 22. 13:50
Checking cluster services and processes
Check the status of the following HACMP daemons:
- The Cluster Manager (clstrmgrES) daemon
- The Cluster Communications (clcomdES) daemon
- The Cluster Information Program (clinfoES) daemon.
When these components are not responding normally, determine if the daemons are active on a cluster node. Use either the options on the SMIT System Management (C-SPOC) > Manage HACMP Services > Show Cluster Services panel or the lssrc command.
For example, to check on the status of all daemons under the control of the SRC, enter:
lssrc -a | grep active
syslogdras 290990active
sendmail mail270484active
portmapportmap286868active
inetd tcpip 295106active
snmpd tcpip 303260active
dpid2 tcpip 299162active
hostmibd tcpip 282812active
aixmibdtcpip 278670active
biodnfs 192646active
rpc.statd nfs 254122 active
rpc.lockd nfs 274584active
qdaemonspooler196720active
writesrv spooler250020active
ctrmc rsct98392 active
clcomdES clcomdES 204920active
IBM.CSMAgentRMrsct_rm90268 active
IBM.ServiceRM rsct_rm229510active
IBM.ERRM rsct_rm188602active
IBM.AuditRMrsct_rm151722active
topsvcstopsvcs602292active
grpsvcsgrpsvcs569376active
emsvcs emsvcs 561188active
emaixosemsvcs 557102active
clstrmgrEScluster544802active
gsclvmd565356active
IBM.HostRMrsct_rm442380activeTo check on the status of all cluster daemons under the control of the SRC, enter: lssrc -g cluster
Note: When you use the -g flag with the lssrc command, the status information does not include the status of subsystems if they are inactive. If you need this information, use the -a flag instead. For more information on the lssrc command, see the man page.
To view additional information on the status of a daemon run the clcheck_server command. The clcheck_server command makes additional checks and retries beyond what is done by lssrc command. For more information, see the clcheck_server man page.
To determine whether the Cluster Manager is running, or if processes started by the Cluster Manager are currently running on a node, use the ps command.
For example, to determine whether the clstrmgrES daemon is running, enter:
ps -ef | grep clstrmgrES
root 18363 3346 3 11:02:05 - 10:20 /usr/es/sbin/cluster/clstrmgrES
root 19028 19559 2 16:20:04 pts/10 0:00 grep clstrmgrES
Checking cluster services and processes
Check the status of the following HACMP daemons:
- The Cluster Manager (clstrmgrES) daemon
- The Cluster Communications (clcomdES) daemon
- The Cluster Information Program (clinfoES) daemon.
When these components are not responding normally, determine if the daemons are active on a cluster node. Use either the options on the SMIT System Management (C-SPOC) > Manage HACMP Services > Show Cluster Services panel or the lssrc command.
For example, to check on the status of all daemons under the control of the SRC, enter:
lssrc -a | grep active
syslogdras 290990active
sendmail mail270484active
portmapportmap286868active
inetd tcpip 295106active
snmpd tcpip 303260active
dpid2 tcpip 299162active
hostmibd tcpip 282812active
aixmibdtcpip 278670active
biodnfs 192646active
rpc.statd nfs 254122 active
rpc.lockd nfs 274584active
qdaemonspooler196720active
writesrv spooler250020active
ctrmc rsct98392 active
clcomdES clcomdES 204920active
IBM.CSMAgentRMrsct_rm90268 active
IBM.ServiceRM rsct_rm229510active
IBM.ERRM rsct_rm188602active
IBM.AuditRMrsct_rm151722active
topsvcstopsvcs602292active
grpsvcsgrpsvcs569376active
emsvcs emsvcs 561188active
emaixosemsvcs 557102active
clstrmgrEScluster544802active
gsclvmd565356active
IBM.HostRMrsct_rm442380activeTo check on the status of all cluster daemons under the control of the SRC, enter: lssrc -g cluster
Note: When you use the -g flag with the lssrc command, the status information does not include the status of subsystems if they are inactive. If you need this information, use the -a flag instead. For more information on the lssrc command, see the man page.
To view additional information on the status of a daemon run the clcheck_server command. The clcheck_server command makes additional checks and retries beyond what is done by lssrc command. For more information, see the clcheck_server man page.
To determine whether the Cluster Manager is running, or if processes started by the Cluster Manager are currently running on a node, use the ps command.
For example, to determine whether the clstrmgrES daemon is running, enter:
ps -ef | grep clstrmgrES
root 18363 3346 3 11:02:05 - 10:20 /usr/es/sbin/cluster/clstrmgrES
Storage2012. 2. 11. 11:04
