[adrotate group=”3,7,8,9″]
We had a lot of issues recently with the hosting LPAR which caused us to un-install all of the Guest partitions except the iOS partition to see if they were having any adverse effects on the Hosting partition. A Reclaim Storage found a number of damaged objects but none of them related to the installs we had done. It also gives us a chance to go back over the installs and document what we did to get them up and running.
So this is a list of the actions we took with a number of pretty pictures to show some of the important steps. The HMC is used to setup the Partitions, we were going to use VPM but the restrictions on the number of Guest partitions meant we had to step up to the HMC/VIOS standard edition to meet the requirements.
This is the current partition information it shows we still have a Hosting partition (SHIELD3) and a Guest partition (IOS2) running. We also have a couple of other servers which are powered off at the moment which are not important for this exercise.
First thing to do is create the configuration for the LPAR, we created one called RedHat with the following requirements.
1. The Partition would use .2 of a processor uncapped.
2. Maximum Virtual processors will be 2.
3. It will have 1GB of memory.
4. It will use the LHEA resource for the LAN.
5. We need a Serial Server Virtual adapter for the console.
6. It will not have Power control
This is the summary which is produced for the LPAR profile.
You will notice we have no Physical IO, all of the storage is provided by the Virtual storage device. The next change we have to make is to add the storage definition to the Hosting partition. We need to add the change twice as the LPAR is currently running because updating the profile will only take effect once the partition is restarted, we need the active partition to be updated to recognize the new storage requirement.
Here is a picture of the updated Host Partitions Virtual adapter settings. We also added it to the profile so when a restart of the LPAR happens the new profile will have the same information. Strange IBM doesn’t automatically update the profile.
That is all we had to do for the HMC configuration we did a quick start of the LPAR with a console screen which showed it connected correctly and we did manage to see the initial startup of the console connection.
Next we created the objects required in the hosting partition to allow the LPAR to be started and installed.
First step is to create the Console user for the remote console access. This allows the console to be started and have sufficient authority to carry out the required tasks. To configure you need to go through SST and create a new Console user which is accessed from the Work with service tools user IDs and Devices menu option then the Service tools user IDs menu option. Most of the default options are acceptable but you need to add at least the following options when you create the new ID.
– System partitions – operations
– System partitions – administration
– Partition remote panel key
Once you have updated the profile you can exit back to the command line and the profile should be enabled for use. Here is a view of the profile authorities after creation.
To install RedHat we need to create a Network Server Description, this is connected to a Network Storage Space which is how the installation can be carried out. To install RedHat we need to create a storage space of about 6GB according to the manuals we looked through. The recent install of Linux SuSe (which was the first install we did) worked well with 10GB of disk space so we thought we would use a bit more for this install just so we can install more utilities so we settled on 20GB.
Creating the NWSD is the first object to be created, there are a couple of items we tripped over so we have added a couple of images of our configs to show them.
First of all we found the use of *AUTO required the Partition entry had to match exactly the HMC definition we had created. We also wanted to control when the partition was started so we changed the Online at IPL to *NO.
The next changes are related to the installation media, we used the WRKLNK command to find the correct path to the install image on the CD, the manuals we had read through previously had totally different paths defined so it is important you find the correct path before you attempt to do the install
We are going to install the product from a *STMF (we were baffled by the AIX install which was installed from CD but had *NWSSTG as the IPL source?). Another gotcha is the Power Control setting, make sure that this is set to *NO for the installation otherwise the NWSD will refuse to vary on. We added the VNC=1 to say we want to have a VNC server started at IPL. This allows us to get to the graphical interface when using RedHat.
Now the definition is created we need to create the network storage space to attach to the definition. As with all IBMi functions you can access them via the menu or list options plus use the command directly. Here is our storage space setup.
After pressing enter the system will go out and allocate the entire space and format it for use. The disk utilization will increase based on the amount allocated not
The format and allocation of the space takes a bit of time so time to make a coffee while you wait. Once the storage space has been created and formatted you now need to link the storage space to the definition.
Now we are ready to activate the partition. You can achieve this by taking option 8 from the WRKNWSD list which will show you the partition and take option 1 against it. If the NWSD does not become ACTIVE you have a configuration problem somewhere which has to be resolved before you can go any further.
We are now ready to install the Linux OS. The redbook we looked at mentioned the ability to use the VNC graphical interface to do the install but as we have no network connection to the install device we could not figure out what they were on about? So we took the console window option as the only way we could install the RHE OS. To get to the console just start the partition from the HMC and select the Open Terminal Window from the Console dropdown. You should see something like the following.
The option 1 which is out of sight in this picture is the first option to take it sets the language for the install. Next we need to set up the boot options. Your install may be different so you may need to change a few things. We were installing from the CD so we had to select the device information using the options presented after pressing option 5. Once the boot device had been set the console will re-connect and see the installation image available on the CD.
We had to press enter a couple of time to get through to the actual installer (anaconda) and we also took the option not to check the media as we decided as it has only just been created and it has not been scratched or anything else it should be OK. The install procedure is pretty straight forward although the selection process for some entries was a bit confusing but after letting the Autopartition do its business and allowing it to remove all existing data the install went off pretty smoothly. We only used the default installation for expediency as we are only going to use this demo development with PHP and Easycom in mind although that did amount to 692 packages..
Once the installation is finished you can remove the installation CD and press enter to reboot the partition. At this point it will not be able to start RedHat because the NWSD is set to start from the CD so it will simple go into the console management software, just close the terminal window at that point.
We now need to tell the NWSD where the real boot device is so we need to go in and set the correct boot point which is the *NWSSTG option. This requires the NWSD is turned off and started again after it is updated. The following screen shows the new settings.
Now we should be able to start the partition again and it will boot into RedHat. For us the partition in the HMC was showing it was loading the Firmware so we took the option to end the partition and started it up again. This did bring up the same state because it was trying to boot from the network, we simply change the boot device back to the Hard Drive we had configured and it started up.
Now we need to configure the network etc so we can access the OS from the VNC client and Putty, as this is specific to our install we have not provided screen shots and this is a normal activity for configuring Linux. Perhaps once we get the Apache Server up and running with Easycom installed we will go over that part?
So that is it we now have a fully functioning RedHat Enterprise partition running as a Guest under an IBMi hosting partition.
If you have any questions about the install or what we have done since let us know.
Chris…