Now that you have your VMControl working and have used it you capture a virtual appliance you might want to move it to other environments, such as importing it into IWD to deploy on any number of environments. The best thing to do is convert it to a ova, which just happens a tar file, just remember AIX tar files are limited to 8GB.
First you will need to log onto the Systems Director and take a note of its unique identifier, this is located as follows -
Left task menu -> System Configuration -> VMControl. Once this is open select the 'Virtual Appliances' tab followed by the 'Image Repository', then just select the relevant one if you have more and you should then see a field called 'Unique Identifier'.
Now we have the identifier from the NIM master, perform the following steps as the root user-
Change to the appliance directory of the virtual appliance that you want to export using the following command:
# cd /export/nim/appliances/<unique identifier>
Copy the .ovf file, the mksysb file found in the directory to a new location.
In the copied .ovf files, change all file reference to use a relative path -
<ovf:File ovf:href="file:///export/nim/appliances/<Unique Identifier>/image1.mksysb" ovf:id="vimRef1" ovf:size="<size>"/>
to:
<ovf:File ovf:href="image1.mksysb" ovf:id="vimRef1" ovf:size="<size>"/>
Now we can create our tar achive ensuring that the ovf is the first file in the archive -
# tar xvf <file-name>.ovf <file-name>.ova
# tar uvf <file-name>.mksysb <file-name>.ova
I've been working with UNIX, predominantly IBM's AIX for a number of years now. From here I hope to pass on some of that information and create a useful resource.
Monday, 28 November 2011
Friday, 25 November 2011
Virtual Appliance Errors
DNZIMN867E Could not exchange SSH key with 172.19.105.25 due to the following error:
2760-287 [dkeyexch] Internal error - exchange script returns unknown error: 1
This means your ssh keys have not been set-up, or you have simply never logged into the HMC from you NIM server to do the ssh exhange, so check with -
# ssh hscroot@cld_hmc1 lssyscfg -r sys -F name
If that fails then take a look at this Developer works page -
https://www.ibm.com/developerworks/wikis/display/WikiPtype/Manual+steps+to+enable+NIM+master+to+HMC+or+IVM+communication
DNZIMN880E Failed to extract NIM Shared Product Object Tree (SPOT) resource from image appliance-0_image-1
There are two possible reasons for this, one is /export/nim filesystem needs more space, or its trying to locate a language file. So check in the Director agents error log here -
# /opt/ibm/director/agent/logs/error-log-0.html
For this message - 0042-001 nim: processing error encountered on "master":
Then along with it you should find something like this - /usr/lib/nls/msg/de_AT not found
/usr/lib/nls/msg/de_LU not found
Just touch these files on the NIM server and you should be able to rerun your deploy with no further occurrences of this error.
DNZIMN356E Could not connect to the image repository
This means that there is a issue talking to the image repository on the NIM from the System Director, which while it might still show up on VMControl all your deploys continue to fail. In some cases this is a issue with what network interface the system is talking over, so if you do have multipule interfaces then you can lock it down as follows -
# vi /opt/ibm/director/data/IPPreference.properties
com.ibm.director.server.preferred.ip=<preferred IP>
com.ibm.director.endpoint.excluded.ip.prefix=<excluded IP>,<excluded IP>
This is not true in all cases, it could relate to firewall issues, or in my case is issue with the levels of the system, so if you run the following on the NIM server to check the install -
# /opt/ibm/director/agent/bin/./lwiupdatemgr.sh -listFeatures|grep nim
com.ibm.director.im.rf.nim.subagent_2.3.1.2 Enabled
So we can see that the nim subagent is enabled, then look to see if the plugin is now active -
# /opt/ibm/director/agent/bin/./lwiplugin.sh -status | grep com.ibm.director.im
214:ACTIVE:Events Plug-in:2.3.1.2:com.ibm.director.im.rf.events
215:RESOLVED:NIM RF NLS Fragment:2.3.1.2:com.ibm.director.im.rf.nim.imaster.nl1
216:ACTIVE:NIM Master Interfaces Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.imaster
217:ACTIVE:NIM Master CAS Agent Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.master
Above is a correctly active agent, if you see the following -
214:ACTIVE:Events Plug-in:2.3.1.2:com.ibm.director.im.rf.events
215:RESOLVED:NIM RF NLS Fragment:2.3.1.2:com.ibm.director.im.rf.nim.imaster.nl1
216:((LAZY)):NIM Master Interfaces Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.imaster
217:INSTALLED:NIM Master CAS Agent Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.master
Then the agent is not installed correctly and it probably means your fix level is to high, in my case when I performed the following updated -
6100-06_AIX_ML:DirectorCommonAgent:6.2.0.1:6.2.0.0:-:AIX 6100-06 Update
6100-06_AIX_ML:DirectorPlatformAgent:6.2.0.1:6.2.0.0:-:AIX 6100-06 Update
From 6.2.0.0 issue above occurred and the only way I could recover the problem was restoring my server from mksysb.
2760-287 [dkeyexch] Internal error - exchange script returns unknown error: 1
This means your ssh keys have not been set-up, or you have simply never logged into the HMC from you NIM server to do the ssh exhange, so check with -
# ssh hscroot@cld_hmc1 lssyscfg -r sys -F name
If that fails then take a look at this Developer works page -
https://www.ibm.com/developerworks/wikis/display/WikiPtype/Manual+steps+to+enable+NIM+master+to+HMC+or+IVM+communication
DNZIMN880E Failed to extract NIM Shared Product Object Tree (SPOT) resource from image appliance-0_image-1
There are two possible reasons for this, one is /export/nim filesystem needs more space, or its trying to locate a language file. So check in the Director agents error log here -
# /opt/ibm/director/agent/logs/error-log-0.html
For this message - 0042-001 nim: processing error encountered on "master":
Then along with it you should find something like this - /usr/lib/nls/msg/de_AT not found
/usr/lib/nls/msg/de_LU not found
Just touch these files on the NIM server and you should be able to rerun your deploy with no further occurrences of this error.
DNZIMN356E Could not connect to the image repository
This means that there is a issue talking to the image repository on the NIM from the System Director, which while it might still show up on VMControl all your deploys continue to fail. In some cases this is a issue with what network interface the system is talking over, so if you do have multipule interfaces then you can lock it down as follows -
# vi /opt/ibm/director/data/IPPreference.properties
com.ibm.director.server.preferred.ip=<preferred IP>
com.ibm.director.endpoint.excluded.ip.prefix=<excluded IP>,<excluded IP>
This is not true in all cases, it could relate to firewall issues, or in my case is issue with the levels of the system, so if you run the following on the NIM server to check the install -
# /opt/ibm/director/agent/bin/./lwiupdatemgr.sh -listFeatures|grep nim
com.ibm.director.im.rf.nim.subagent_2.3.1.2 Enabled
So we can see that the nim subagent is enabled, then look to see if the plugin is now active -
# /opt/ibm/director/agent/bin/./lwiplugin.sh -status | grep com.ibm.director.im
214:ACTIVE:Events Plug-in:2.3.1.2:com.ibm.director.im.rf.events
215:RESOLVED:NIM RF NLS Fragment:2.3.1.2:com.ibm.director.im.rf.nim.imaster.nl1
216:ACTIVE:NIM Master Interfaces Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.imaster
217:ACTIVE:NIM Master CAS Agent Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.master
Above is a correctly active agent, if you see the following -
214:ACTIVE:Events Plug-in:2.3.1.2:com.ibm.director.im.rf.events
215:RESOLVED:NIM RF NLS Fragment:2.3.1.2:com.ibm.director.im.rf.nim.imaster.nl1
216:((LAZY)):NIM Master Interfaces Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.imaster
217:INSTALLED:NIM Master CAS Agent Plug-in:2.3.1.2:com.ibm.director.im.rf.nim.master
Then the agent is not installed correctly and it probably means your fix level is to high, in my case when I performed the following updated -
6100-06_AIX_ML:DirectorCommonAgent:6.2.0.1:6.2.0.0:-:AIX 6100-06 Update
6100-06_AIX_ML:DirectorPlatformAgent:6.2.0.1:6.2.0.0:-:AIX 6100-06 Update
From 6.2.0.0 issue above occurred and the only way I could recover the problem was restoring my server from mksysb.
Wednesday, 16 November 2011
POWER Cloud Build Notes
- Requirements -
- IBM Systems Director with VMControl -
1/2 Procs Ent1 - Uncapped - 2/4 Virt - 4Gb Mem - 4-8Gb storage, extra if it is to be the NIM Master too. OS 5.3 - 6.1 - 7.1
NIM Server requires Director Common Agent installed if it is not on the same node + VMControl sub agent
POWER System 6/7
HMC
Integrated Virtualization Manager (IVM)
NOTES -
After a initial install I hit a number of problem getting the director server to install NIM agents on the NIM server and other issues such as complete System Director failure after it not being used for a week. In addition there where further problems getting the VMControl to work and Director rediscovering and inventorying known systems correctly. This revision should resolve those issues and get a system built ready to be integrated in any Cloud environment.
It should be stated that this is only the second revision, so some details have been take from memory is searches of history files from when I progressing round a problem. I have plans for further systems and more test when more equipment arrives.
I will also be looking into a possible issue that I can't get VMControl working correctly due to the HMC and System Director/NIM Server being on different VLANs, the plan being to see if I get the same failure when the HMC is moved to the same VLAN as everything else.
- VIO/System Director server install -
First you will need to get a VIO server installed - normal details can be followed to do this. Once its up and running and you can created a number of Virtual adapters to use you will need to create a partion for the systems director to be installed on using the specifications above, so we need some system pools created -
- Good VIO quick sheet - http://www.tablespace.net/quicksheet/powervm-quickstart.html
First lets create a boot volume group on the VIO Server, this is where we will keep all the boot/rootvg data, then followed by a data VG.
Create boot storage pool - # mksp cloud_boot hdisk1
Then the data pool -
# mksp cloud_data hdisk2 hdisk3
Note- the disks for the storage pool can be anything you like, either from the POWER systems internal disk or SAN attached, its your choice depending on what is available. In this case its all internal disk.
Then change the default storage pool from rootvg to cloud_boot -
# chsp -default cloud_boot
list those pools -
# lssp
Pool Size(mb) Free(mb) Alloc Size(mb) BDs Type
rootvg 69888 46848 128 0 LVPOOL
cloud_boot 69888 69888 64 0 LVPOOL
cloud_data 139776 139776 64 0 LVPOOL
Create a client disk for the System Director on adapter vhost0 from cloud_boot storage pool -
# mkbdsp -sp cloud_boot 20G -bd lv_pclm1_boot -vadapter vhost0
This server will also be a NIM server so we will need to create some data storage for it -
# mkbdsp -sp cloud_data 40G -bd lv_pclm1_data -vadapter vhost0
Note- keep the lv names no longer the 15 characters, else it will not map the disks correct to the vhost as is passes the command over, so 'lsmap -all' fails.
To list the details in the storage pool -
# lssp -bd - shows the default pool, in this case cloud_boot
# lssp -bd -sp cloud_data - to list other pools.
Once these pools are created and assigned over to the server you can get it booted for the network install, at this stage you need to boot into SMS and select the network device and configure the ip details.
- System Director Install -
Obtain the latest System Director code, this example is 6.2.1.2 -
Extract the gzip/tar file into a relevant dir with enough space, about 3Gb in all. Source file is - "SysDir6_2_Server_AIX.tar"
Kick of the install with ./dirinstall.server
It should automatically expand the /opt filesystem to ensure that there is sufficient space for the install.
For proper environment setup for the database on this platform, please go to the IBM Systems Director Info Center:
http://publib.boulder.ibm.com/infocenter/director/v6r2x/topic/com.ibm.director.install.helps.doc/fqm0_t_install_config_database_application.html
You must configure the agent manager prior to starting the server.
To configure the agent manager, run # /opt/ibm/director/bin/configAgtMgr.sh
This will allow you to set the System Director user (root) and the default passwords, in our working environment they all match that of 'root'.
To start the server manually, run
# /opt/ibm/director/bin/smstart (this path can be added to the .profile PATH variable)
Monitor the status of the start with the following command as it can take some time to start -
# smstatus -r
Ctrl-c breaks out once the status changes to active.
- Install the IBM Systems Director VMControl 2.3.1 Plugin -
Confirm that IBM System Director 6.2.1.2 is installed on your IBM Systems Director server.
Open a browser to https://<director>:8422/ibm/console and log on.
On the Welcome to IBM Systems Director panel, click on the Manage tab. Confirm IBM Systems Director 6.2.1.2 is installed
Logout of IBM Systems Director.
Start a telnet (or ssh) session to your IBM Systems Director server and sign on.
Create a directory for installing VMControl Enterprise Edition 2.3.1 and then change directories to it. # mkdir /tmp/sd
# cd /tmp/sd
The IBM System Director VMControl Enterprise Edition 2.3.1 for AIX 60 day trial can be downloaded at ->
http://www.ibm.com/systems/software/director/vmcontrol/enterprise.html
Unzip and decompress the IBM System Director VMControl Enterprise Edition 2.3.1 installation file # gunzip –c /tmp/sd/SysDir_VMControl_2_3_1_Linux_AIX.tar.gz | tar xvf –
Using VI from PuTTY or other editor, edit the /tmp/sd/installer.properties file as follows and save it.
Installer_UI = silent
License _accepted = true
Install IBM System Director VMControl Enterprise Edition 2.3.1 as follows:
# /tmp/sd/IBMSystemsDirector-VMControl-Setup.sh –i silent
When completed successfully, stop and restart IBM System Director
# /opt/ibm/director/bin/smstop
# /opt/ibm/director/bin/smstart
# /opt/ibm/director/bin/smstatus –r
Once IBM Systems Director is active, point your browser to https://<director>:8422/ibm/console and log on.
On the Welcome to IBM Systems Director panel, click on the Manage tab.
You should find that VMControl 2.3.1 is now listed.
- Part 2: Installing the VMControl 2.3.1.2 fix pack -
Start a telnet (or ssh) session to your IBM Systems Director server and sign on.
Log into the System Director and create the following dir for the download to be installed to. # mkdir /tmp/sd/vmc
# cd /tmp/sd/vmc
This fixpack can be downloaded from IBM Fix Central at -> http://www-933.ibm.com/support/fixcentral/ - Package should be com.ibm.director.vmc.collection_2.3.1.2
Confirm that the VMControl 2.3.1.2 fix pack files are located in the /tmp/sd/vmc directory.
I have tried at this point installing this on the local server using the graphical interface, but so far it fails. I have found that I need to run a up-to-date dicovery of the system - # smcli collectinv -p "All Inventory" -n <node_name>
Once this has completed you can then update the system with the VMControl package.
# smcli importupd -vr /tmp/sd/vmc
# smcli installupd -v -n pclm1_sdvmc -u com.ibm.director.vmc.collection_2.3.1.2
When completed successfully, stop and restart IBM System Director
# /opt/ibm/director/bin/smstop
# /opt/ibm/director/bin/smstart
# /opt/ibm/director/bin/smstatus –r
When IBM Systems Director is active, point your browser to https://<director>:8422/ibm/console and log on.
Confirm that VMControl is now at 2.3.1.2 via the Welcome to Systems Director/Manage tab
- NIM Master install -
Build the LPAR you are going to install the NIM master on either a standalone LPAR or in this case the Service Director LPAR -
Install the AIX OS release that you want to the NIM master to serve, remembering that the NIM server has to be at the current or higher OS/patch level then the clients it serves.
Make sure you are happy with the network setup of the NIM master, speed and such, so you can get things set-up. Remember you can always make changes later, but it is easier to ensure this is all correct first.
Install the following filesets if they are not installed already:
dsm.core
openssh.base.client
openssl.base
bos.sysmgt.nim.master
bos.sysmgt.nim.spot
bos.sysmgt.nim.client
bos.net.nfs.server
Now we can create the NIM master with the following script -
# /usr/sbin/nim_master_setup -a mk_resource=yes -a volume_group=nimvg -a device=/mnt/610TL06_lpp_res
In the example above I had mounted a known good lpp source from another nim server and used that, but you can use mksysb's or a DVD image (default).
# lsnim - list the now created NIM resources.
From here you can then add other LPP sources and SPOTs but at the moment we want to make sure that the System Director picks up the NIM respository and installs the NIM agent. So now you have created the NIM filesystem it would be best to increase the size to about 30-40Gig.
# chfs -a size=30G /export/nim
Then perform a inventory on the Director server to pick up the changes to the server.
# smcli collectinv -p "All Inventory" -n <node>
- Installing the VMControl NIM subagent -
In the IBM Systems Director navigation pane, expand Systems Configuration and then click VMControl -
Notice message DNZIMC7081 – No image repositories were detected in IBM Systems Director. Click Install Agents.
On the Welcome to Agent Installation Wizard panel, click Next.
Click on Common Agent Subagent Packages.
Select CommonAgentSubagent_VMControl_NIM-2.3.1.2 and click Add then click Next.
Select your IBM Systems Director server and click Add. And then click Next.
Review the summary of the Install Agent Task and then click Finish.
Run the job now and display its properties. Monitor the job for a successful completion.
Collect inventory on your IBM Systems Director (and NIM master) server
When complete, in the IBM Systems Director navigation pane, expand Systems Configuration and then click VMControl.
Ensure that message DNZIMC709I is now displayed indicating that No virtual appliances were found.
Then ensure that you restart the Director service -
# smstop;smstart;smstatus -r
- NIM Server further configuration -
We can now update the bosinst_data resource so that we have hands free installation, we copy a file from the master system to the NIM volume group and define it. This time though, we also modify the resource. Read carefully!
# cp /bosinst.data /export/nim/bosinst_data/6100-05bid_ow
This was the NIM resource - lsnim -l 6100-05bid_ow that the NIM creation script created.
For an automatic installation there are a number of lines we need to change:
The lines to change are from:
PROMPT = yes
RECOVER_DEVICES = Default
ACCEPT_LICENSES =
ACCEPT_SWMA =
IMPORT_USER_VGS =
CONSOLE = Default
CREATE_JFS2_FS = Default
to: PROMPT = no
RECOVER_DEVICES = no
ACCEPT_LICENSES = yes
ACCEPT_SWMA = yes
IMPORT_USER_VGS = no
CONSOLE = /dev/vty0
CREATE_JFS2_FS = yes
add
ENABLE_64BIT_KERNEL = yes
And simplify the target_disk_data: stanza from:
target_disk_data:
PVID = 00f67207e26dbb8d
PHYSICAL_LOCATION = U8233.E8B.107207P-V41-C21-T1-L8100000000000000
CONNECTION = vscsi0//810000000000
LOCATION =
SIZE_MB = 70006
HDISKNAME = hdisk0
to: target_disk_data:
HDISKNAME = hdisk0
Remember it is the file /export/nim/bosinst_data/6100-05bid_ow that needs to be edited, though you can of course change any more lines that you might want.
If you want to create and add any other additional resources then the following commands should help -
First this a package check to pick out any duplicate files in any lpp sources -
# /usr/lib/instl/lppmgr -d /export/nim/lpp_source/<lpp_source> -u -b -x -r
From a lpp source you can then create a SPOT, in this case from lpp 610TL06-lpp into a spot called 610TL06_spot.
# nim -o define -t spot -a server=master -a source=610TL06-lpp -a location=/export/nim/spot/610TL06_spot 610TL06_spot
To perform a command line discovery -
# smcli discover -i <ip-address>,<ip-address>
Then you should be able to perfrom a install via the NIM server
# nim -o bos_inst -a lpp_source=<lpp-source> -a spot=<spot> -a bosinst_data=<bos-inst> -a no_client_boot=no <server-name>
- IBM Systems Director with VMControl -
1/2 Procs Ent1 - Uncapped - 2/4 Virt - 4Gb Mem - 4-8Gb storage, extra if it is to be the NIM Master too. OS 5.3 - 6.1 - 7.1
NIM Server requires Director Common Agent installed if it is not on the same node + VMControl sub agent
POWER System 6/7
HMC
Integrated Virtualization Manager (IVM)
NOTES -
After a initial install I hit a number of problem getting the director server to install NIM agents on the NIM server and other issues such as complete System Director failure after it not being used for a week. In addition there where further problems getting the VMControl to work and Director rediscovering and inventorying known systems correctly. This revision should resolve those issues and get a system built ready to be integrated in any Cloud environment.
It should be stated that this is only the second revision, so some details have been take from memory is searches of history files from when I progressing round a problem. I have plans for further systems and more test when more equipment arrives.
I will also be looking into a possible issue that I can't get VMControl working correctly due to the HMC and System Director/NIM Server being on different VLANs, the plan being to see if I get the same failure when the HMC is moved to the same VLAN as everything else.
- VIO/System Director server install -
First you will need to get a VIO server installed - normal details can be followed to do this. Once its up and running and you can created a number of Virtual adapters to use you will need to create a partion for the systems director to be installed on using the specifications above, so we need some system pools created -
- Good VIO quick sheet - http://www.tablespace.net/quicksheet/powervm-quickstart.html
First lets create a boot volume group on the VIO Server, this is where we will keep all the boot/rootvg data, then followed by a data VG.
Create boot storage pool - # mksp cloud_boot hdisk1
Then the data pool -
# mksp cloud_data hdisk2 hdisk3
Note- the disks for the storage pool can be anything you like, either from the POWER systems internal disk or SAN attached, its your choice depending on what is available. In this case its all internal disk.
Then change the default storage pool from rootvg to cloud_boot -
# chsp -default cloud_boot
list those pools -
# lssp
Pool Size(mb) Free(mb) Alloc Size(mb) BDs Type
rootvg 69888 46848 128 0 LVPOOL
cloud_boot 69888 69888 64 0 LVPOOL
cloud_data 139776 139776 64 0 LVPOOL
Create a client disk for the System Director on adapter vhost0 from cloud_boot storage pool -
# mkbdsp -sp cloud_boot 20G -bd lv_pclm1_boot -vadapter vhost0
This server will also be a NIM server so we will need to create some data storage for it -
# mkbdsp -sp cloud_data 40G -bd lv_pclm1_data -vadapter vhost0
Note- keep the lv names no longer the 15 characters, else it will not map the disks correct to the vhost as is passes the command over, so 'lsmap -all' fails.
To list the details in the storage pool -
# lssp -bd - shows the default pool, in this case cloud_boot
# lssp -bd -sp cloud_data - to list other pools.
Once these pools are created and assigned over to the server you can get it booted for the network install, at this stage you need to boot into SMS and select the network device and configure the ip details.
- System Director Install -
Obtain the latest System Director code, this example is 6.2.1.2 -
Extract the gzip/tar file into a relevant dir with enough space, about 3Gb in all. Source file is - "SysDir6_2_Server_AIX.tar"
Kick of the install with ./dirinstall.server
It should automatically expand the /opt filesystem to ensure that there is sufficient space for the install.
For proper environment setup for the database on this platform, please go to the IBM Systems Director Info Center:
http://publib.boulder.ibm.com/infocenter/director/v6r2x/topic/com.ibm.director.install.helps.doc/fqm0_t_install_config_database_application.html
You must configure the agent manager prior to starting the server.
To configure the agent manager, run # /opt/ibm/director/bin/configAgtMgr.sh
This will allow you to set the System Director user (root) and the default passwords, in our working environment they all match that of 'root'.
To start the server manually, run
# /opt/ibm/director/bin/smstart (this path can be added to the .profile PATH variable)
Monitor the status of the start with the following command as it can take some time to start -
# smstatus -r
Ctrl-c breaks out once the status changes to active.
- Install the IBM Systems Director VMControl 2.3.1 Plugin -
Confirm that IBM System Director 6.2.1.2 is installed on your IBM Systems Director server.
Open a browser to https://<director>:8422/ibm/console and log on.
On the Welcome to IBM Systems Director panel, click on the Manage tab. Confirm IBM Systems Director 6.2.1.2 is installed
Logout of IBM Systems Director.
Start a telnet (or ssh) session to your IBM Systems Director server and sign on.
Create a directory for installing VMControl Enterprise Edition 2.3.1 and then change directories to it. # mkdir /tmp/sd
# cd /tmp/sd
The IBM System Director VMControl Enterprise Edition 2.3.1 for AIX 60 day trial can be downloaded at ->
http://www.ibm.com/systems/software/director/vmcontrol/enterprise.html
Unzip and decompress the IBM System Director VMControl Enterprise Edition 2.3.1 installation file # gunzip –c /tmp/sd/SysDir_VMControl_2_3_1_Linux_AIX.tar.gz | tar xvf –
Using VI from PuTTY or other editor, edit the /tmp/sd/installer.properties file as follows and save it.
Installer_UI = silent
License _accepted = true
Install IBM System Director VMControl Enterprise Edition 2.3.1 as follows:
# /tmp/sd/IBMSystemsDirector-VMControl-Setup.sh –i silent
When completed successfully, stop and restart IBM System Director
# /opt/ibm/director/bin/smstop
# /opt/ibm/director/bin/smstart
# /opt/ibm/director/bin/smstatus –r
Once IBM Systems Director is active, point your browser to https://<director>:8422/ibm/console and log on.
On the Welcome to IBM Systems Director panel, click on the Manage tab.
You should find that VMControl 2.3.1 is now listed.
- Part 2: Installing the VMControl 2.3.1.2 fix pack -
Start a telnet (or ssh) session to your IBM Systems Director server and sign on.
Log into the System Director and create the following dir for the download to be installed to. # mkdir /tmp/sd/vmc
# cd /tmp/sd/vmc
This fixpack can be downloaded from IBM Fix Central at -> http://www-933.ibm.com/support/fixcentral/ - Package should be com.ibm.director.vmc.collection_2.3.1.2
Confirm that the VMControl 2.3.1.2 fix pack files are located in the /tmp/sd/vmc directory.
I have tried at this point installing this on the local server using the graphical interface, but so far it fails. I have found that I need to run a up-to-date dicovery of the system - # smcli collectinv -p "All Inventory" -n <node_name>
Once this has completed you can then update the system with the VMControl package.
# smcli importupd -vr /tmp/sd/vmc
# smcli installupd -v -n pclm1_sdvmc -u com.ibm.director.vmc.collection_2.3.1.2
When completed successfully, stop and restart IBM System Director
# /opt/ibm/director/bin/smstop
# /opt/ibm/director/bin/smstart
# /opt/ibm/director/bin/smstatus –r
When IBM Systems Director is active, point your browser to https://<director>:8422/ibm/console and log on.
Confirm that VMControl is now at 2.3.1.2 via the Welcome to Systems Director/Manage tab
- NIM Master install -
Build the LPAR you are going to install the NIM master on either a standalone LPAR or in this case the Service Director LPAR -
Install the AIX OS release that you want to the NIM master to serve, remembering that the NIM server has to be at the current or higher OS/patch level then the clients it serves.
Make sure you are happy with the network setup of the NIM master, speed and such, so you can get things set-up. Remember you can always make changes later, but it is easier to ensure this is all correct first.
Install the following filesets if they are not installed already:
dsm.core
openssh.base.client
openssl.base
bos.sysmgt.nim.master
bos.sysmgt.nim.spot
bos.sysmgt.nim.client
bos.net.nfs.server
Now we can create the NIM master with the following script -
# /usr/sbin/nim_master_setup -a mk_resource=yes -a volume_group=nimvg -a device=/mnt/610TL06_lpp_res
In the example above I had mounted a known good lpp source from another nim server and used that, but you can use mksysb's or a DVD image (default).
# lsnim - list the now created NIM resources.
From here you can then add other LPP sources and SPOTs but at the moment we want to make sure that the System Director picks up the NIM respository and installs the NIM agent. So now you have created the NIM filesystem it would be best to increase the size to about 30-40Gig.
# chfs -a size=30G /export/nim
Then perform a inventory on the Director server to pick up the changes to the server.
# smcli collectinv -p "All Inventory" -n <node>
- Installing the VMControl NIM subagent -
In the IBM Systems Director navigation pane, expand Systems Configuration and then click VMControl -
Notice message DNZIMC7081 – No image repositories were detected in IBM Systems Director. Click Install Agents.
On the Welcome to Agent Installation Wizard panel, click Next.
Click on Common Agent Subagent Packages.
Select CommonAgentSubagent_VMControl_NIM-2.3.1.2 and click Add then click Next.
Select your IBM Systems Director server and click Add. And then click Next.
Review the summary of the Install Agent Task and then click Finish.
Run the job now and display its properties. Monitor the job for a successful completion.
Collect inventory on your IBM Systems Director (and NIM master) server
When complete, in the IBM Systems Director navigation pane, expand Systems Configuration and then click VMControl.
Ensure that message DNZIMC709I is now displayed indicating that No virtual appliances were found.
Then ensure that you restart the Director service -
# smstop;smstart;smstatus -r
- NIM Server further configuration -
We can now update the bosinst_data resource so that we have hands free installation, we copy a file from the master system to the NIM volume group and define it. This time though, we also modify the resource. Read carefully!
# cp /bosinst.data /export/nim/bosinst_data/6100-05bid_ow
This was the NIM resource - lsnim -l 6100-05bid_ow that the NIM creation script created.
For an automatic installation there are a number of lines we need to change:
The lines to change are from:
PROMPT = yes
RECOVER_DEVICES = Default
ACCEPT_LICENSES =
ACCEPT_SWMA =
IMPORT_USER_VGS =
CONSOLE = Default
CREATE_JFS2_FS = Default
to: PROMPT = no
RECOVER_DEVICES = no
ACCEPT_LICENSES = yes
ACCEPT_SWMA = yes
IMPORT_USER_VGS = no
CONSOLE = /dev/vty0
CREATE_JFS2_FS = yes
add
ENABLE_64BIT_KERNEL = yes
And simplify the target_disk_data: stanza from:
target_disk_data:
PVID = 00f67207e26dbb8d
PHYSICAL_LOCATION = U8233.E8B.107207P-V41-C21-T1-L8100000000000000
CONNECTION = vscsi0//810000000000
LOCATION =
SIZE_MB = 70006
HDISKNAME = hdisk0
to: target_disk_data:
HDISKNAME = hdisk0
Remember it is the file /export/nim/bosinst_data/6100-05bid_ow that needs to be edited, though you can of course change any more lines that you might want.
If you want to create and add any other additional resources then the following commands should help -
First this a package check to pick out any duplicate files in any lpp sources -
# /usr/lib/instl/lppmgr -d /export/nim/lpp_source/<lpp_source> -u -b -x -r
From a lpp source you can then create a SPOT, in this case from lpp 610TL06-lpp into a spot called 610TL06_spot.
# nim -o define -t spot -a server=master -a source=610TL06-lpp -a location=/export/nim/spot/610TL06_spot 610TL06_spot
To perform a command line discovery -
# smcli discover -i <ip-address>,<ip-address>
Then you should be able to perfrom a install via the NIM server
# nim -o bos_inst -a lpp_source=<lpp-source> -a spot=<spot> -a bosinst_data=<bos-inst> -a no_client_boot=no <server-name>
Subscribe to:
Posts (Atom)