As much as IBM Systems Directors once setup up is a great tool for managing your IBM Power systems, it in that initial set-up period is a complete pain to debug. I as much as anyone else completely hate it when its not working. As its the base virtualisation framework for Cloud on Power I've spend a lot of time with it over the years and there always seems to be a new problem and error that I've not seen before. So just as I continue to work on issues, I'm going to try to continue to log the errors and fixes that I've seen and made. I already have some entries on my blog from issues I've seen before so here are the links to those:
And at the same time you might want to take a look at these as depending on your systems they could pop up too:
Flex - After VIO Password and IP change 'Request Access' fails (I know it says Flex, but your see the same issue with ISD)
Now the most important file that your need to be taking a look at is this one:
This is when Director stores those all important details about the errors messages is passes when you've been running commands and task either via the command line or GUI. In there is also a number of other important files:
Your notice some files call '.save#', as this is just a text file the system populates and it can get rather busy I tend to archive the information off and flush the file to make it easier to work out which errors are current. For example in the issue I'm working on at the moment, everytime I run collections against a VIOS server it completed with a error, so I copy the current error-log-0.html to error-log-0.html.save# then flush the file with a redirect > /opt/ibm/director/lwi/logs/error-log-0.html and then run collections. If you've look through this log yourself your understand how much help this is! At the same time I've tried to reduce the number servers I've investigating, so this issues was on all my VIOS servers so to reduce the logging its best to action once at time. The information can run into multiple pages and just become confusing to look at.
Collection of Inventory continuously completes with DNZCLI0621E error
Over the last month I decided to update and rebuild my SmartCloud Entry deployment system, I built it at the IIC in Hursley so that people could quickly deploy IBM Power based images, either AIX, iSeries or Linux quickly without having to enable me to do the work. Not perfect for full on performance based engagements or specific hardware/software scenarios but a good place to do testing, porting, education and more. As part of the update I rebuild the SAN and IBM Systems Director, along with updated to the VIO Servers and HMC to enable me to support PowerVC and PowerVP. After the updates and setup of the SAN import of the systems and creations of the data sources for VMControl and Storage Control I hit a issue in that I couldn't import any new virtual appliances. I could see the storage and all the SAN mapped out okay, but nothing worked. Looking at the logs from the import it would seem that my VIO Servers had not had a compete and full inventory, so I ran a collection on them and it produced this error:
Looking in the error-log-0.html it would seem that the VIO Servers I have are down level on the sub agents, causing problems with collections:
You can download the latest version of the related agents for all systems on the IBM Systems Director Agents Download page, just follow the embedded link. Once you've downloaded the agents that you need this entry in the Infocentre covers the install of the packages, either with installp or scripted.