A few days ago I finally got the answer why my VIO Servers cannot be updated through IBM Systems Director. It took my one day and a half of work with an IBM Systems Director Guru to find a solution. A day later I have to update 4 old VIO Servers from 18.104.22.168-FP-23 to 22.214.171.124, indeed the challenge was to update these olds VIO Servers with Systems Director. Here are the steps performed, problems, and solutions that appears while updating these VIO Servers.
Manual Agent installation, discovery, access, inventory
VIO Server 126.96.36.199-FP-23 is an old VIO Server release (30 April 2010) not supported by Systems Director, so Common agent has to be installed manually. Common agent can be installed as root but as to be run as padmin user; if Common agent is running as root Systems Director can’t connect to the agent.
Copy Common agent for ISD to VIO Server and install Common Agent
Common agent can be found on Systems Director host and can be easily copied to VIO Server :
- Copying Common agent to VIO Server, default directory for Common agent installer is : /opt/ibm/director/packaging/agent/common/aix/ :
# scp /opt/ibm/director/packaging/agent/common/aix/6.3/dir6_3_commonagent_aix padmin@vio2:/tmp
# oem_setup_env # sh /tmp/dir6_3_commonagent_aix ............................................................................... dir6_3_commonagent_aix self-extracting installation program... Please wait... /tmp/DirectorAgentselfextract.9044022/dirinstall.agent: test: argument expected +=============================================================================+ Start of product installation on vios1 +=============================================================================+ /var free space not enough, less than 512M, trying to increase it...[done] [..] Installation of IBM Systems DirectorPlatformAgent completed successfully.
Check CAS protocol
A working Common Agent with CAS protocol discovered by Systems Director is a prerequisite to push a VIO Server updates through Systems Director. Check CAS protocol is discovered by System director :
- Discovering :
# smcli discover -H vios1
# smcli accesssys -u padmin vios1
# smcli collectinv -p "All Inventory" vios1
# smcli lssys -A Protocols vios1
Communication with Common agent can fail because System Director is demanding some prerequisites. I my case Common agent was not working because of these problems :
Common agent running as root
Check Common agent is running as padmin and not as root user, and is running at VIO Server boot :
# cfgsvc DIRECTOR_agent -attr RESTART_ON_REBOOT=true # startsvc DIRECTOR_agent Starting cimserver... # lssvc DIRECTOR_agent RESTART_ON_REBOOT:TRUE # ps -ef |grep CIM padmin 6684914 5505090 0 12:51:57 pts/0 0:00 grep CIM padmin 7602396 1 0 12:51:32 - 0:00 /opt/freeware/cimom/pegasus/bin/CIM_diagd
Timezone on Systems Director and VIO Server are identical
Check timezone are identical on VIO Server and on Systems Director (in my case Europe/Paris):
- VIO Server :
# chdate -timezone Europe/Paris # oem_setup_env # grep ^TZ /etc/environment TZ=Europe/Paris
# grep ^TZ /etc/environment TZ=Europe/Paris
Loopback/Localhost is present on VIO Server
If -like me- your are updating and old VIO Server check loopback is present in /etc/hosts file. It’s a well know problem identify by IBM. http://www-01.ibm.com/support/docview.wss?uid=isg3T1012169
# oem_setup_env # hostent -a 127.0.0.0 -h "loopback localhost" # hostent -a ::1 -h "loopback localhost" # grep loopback /etc/hosts
I highly recommend to setup your DNS resolution correctly. Systems Director has to resolve (resolution and inverse resolution) all its clients.Systems Director‘s clients have to resolve (resolution and inverse resolution) the Director itself :
- vios1 has to resolve Systems Director :
# oem_setup_env # host -n -t A sysdir.domain.test sysdir.mydomain.test has address 10.10.122.104 # host -n -t PTR 10.10.122.104 188.8.131.52.IN-ADDR.ARPA domain name pointer sysdir.mydomain.test # tail -1 /etc/netsvc.conf hosts = local4, bind4
# host -n -t A vios1.mydomain.test vios1.mydomain.test has address 10.10.122.61 # host -n -t PTR 10.10.122.61 184.108.40.206.IN-ADDR.ARPA domain name pointer vios1.mydomain.test
Unfortunately IBM Systems Director cannot remove efix. You have to remove theses fixes manually before any update try. Hope this can be corrected by IBM.
- List installed efixes to remove it later :
# oem_setup_env # emgr -l ID STATE LABEL INSTALL TIME UPDATED BY ABSTRACT === ===== ========== ================= ========== ====================================== 1 S IV16920s02 03/21/12 14:28:52 MSNENTDD LINK DOWN AFTER REBOOT. [..]
# emgr -r -L IV16920s02
I will not describe the update phase. You can check part 2 of this serie here. Here are a series of screenshots and output while updating systems :
- System Director updating from 220.127.116.11 to 18.104.22.168 :
- System Director updating from 22.214.171.124-FP-23 to 126.96.36.199 (note thay System Director has to pass two updates in this case, 188.8.131.52-FP-23 to 184.108.40.206, then 220.127.116.11 to 18.104.22.168
[..] October 3, 2012 10:44:20 AM CEST-Level:150-MEID:6248--MSG: ATKUPD632I The Installation Staging task is starting to process system "vio2". October 3, 2012 10:44:20 AM CEST-Level:150-MEID:6248--MSG: ATKUPD911I The updates will be copied to the installation staging server "vio2" to path "/opt/VIOS/update". October 3, 2012 10:57:36 AM CEST-Level:150-MEID:6248--MSG: ATKUPD686I The update "VIOS_22.214.171.124-FP24-SP03" has been staged for installation to "/opt/VIOS/update/VIOS/VIOS_126.96.36.199-FP24-SP03/" on the managed resource "vio2" successfully. October 3, 2012 11:25:59 AM CEST-Level:150-MEID:6248--MSG: ATKUPD686I The update "VIOS_188.8.131.52-FP24" has been staged for installation to "/opt/VIOS/update/VIOS/VIOS_184.108.40.206-FP24/" on the managed resource "vio2" successfully. October 3, 2012 11:25:59 AM CEST-Level:150-MEID:6248--MSG: ATKUPD633I The Installation Staging task has finished processing system "vio2". October 3, 2012 11:25:59 AM CEST-Level:150-MEID:0--MSG: ATKUPD630I The update installation staging has completed. October 3, 2012 11:25:59 AM CEST-Level:150-MEID:6248--MSG: ATKUPD760I Start processing update "VIOS_220.127.116.11-FP24" and system "vio2". October 3, 2012 1:31:26 PM CEST-Level:150-MEID:6248--MSG: ATKUPD764I Update "VIOS_18.104.22.168-FP24" was installed on system "vio2" successfully. October 3, 2012 1:31:26 PM CEST-Level:150-MEID:6248--MSG: ATKUPD760I Start processing update "VIOS_22.214.171.124-FP24-SP03" and system "vio2". October 3, 2012 2:36:47 PM CEST-Level:150-MEID:6248--MSG: ATKUPD764I Update "VIOS_126.96.36.199-FP24-SP03" was installed on system "vio2" successfully. October 3, 2012 2:37:20 PM CEST-Level:100-MEID:223519--MSG: vio2 client job status changed to "Complete". [..] October 3, 2012 2:43:00 PM CEST-Level:150-MEID:0--MSG: ATKUPD727I The update install task has finished successfully.
While updating VIO Server from 188.8.131.52-FP-23 to 184.108.40.206 Systems Director wasn’t able to update in 220.127.116.11. I have to update from 18.104.22.168 to 22.214.171.124 trough the old fashion way with updateios command. If anybody can give me an answer on this point it’ll be appreciated.
Hope this can help.