OSG Operations Training Session April 16th, 2012
OSG Operations will be hosting a monthly training for the OSG Operations Support staff. The goal of these session are to increase the knowledge of the staff to triage and troubleshoot common issues.
This will be the first training session and will focus on the VDT and OSG Software repositories, installations, and common software issues.
Alain Roy, Kyle Gross, Scott Teige, Elizabeth Chism, Chris Pipes, Soichi Hayashi, Alain Deximo, Tom Lee and Rob Quick.
- 9:00 - 10:00 OSG Software Training
- This is be a Powerpoint presentation. I will walk through what software is installed on a the CE, how it is installed, and what the software does in a fair amount of detail. For example:
- How do we install with RPMs?
- How does a job arrive at a CE? (gatekeeper & job manager daemons)
- How do we authenticate?
- What's different in the new OSG 3.0 (as compared to OSG 1.2?)
- How do Gratia probes work?
- How does CEMon/GIP work?
- There will be lots of details, not just hand-waving. Process names, file locations, etc...
- 10:00 - 10:30 A live demonstration of installing the CE
- Alain will do an installation of the CE live. He'll run a couple of jobs and show what happens (Gratia data, GIP data, logs...)
- 10:30 - 10:45 Coffee break
- 10:45 - 11:15 Continue live demo
- 2:00 - 3:00 Alain has to run the OSG Software Team Meeting
- 3:15 - 5:00 Open discussion, aka Stump the expert
- Ask Alain any questions you like. We'll work through any questions people have.
- Walk through specific GOC tickets if desired
While it is hard to gauge success of the training short term, all initial feedback has been positive. I've asked the support staff to link tickets that they use knowledge of OSG Software Training session to Appendix A of this page.
This will be an informal list of tickets that staff referenced the knowledge used in this training. This list may not be exhaustive and will be cut off on May 31st, 2012.
- Kyle Gross - This presentation definitely helped me have a better understanding of how to look at things with regard to the OSG Software Stack. I learned some new little tips and tricks to help me investigate problems. Alain's knowledge of the software stack is honestly staggering. I have been able to better understand some tickets that have been handed to the Software Team now. I look forward to getting some tickets that I can handle on my own. I think this training was a success and set a high bar for the trainings to follow.
- Elizabeth Chism - The presentation was good and useful. I ironed out the problems on my CE. I would have liked to have dug deeper into some other hands on stuff, but there were varying levels of preparedness among the group for such demonstrations. I have already noticed improvements to our support practices so I would consider this a resounding success. Alain Roy was able to answer my questions and increase my understanding of the software.
- Alain Deximo - I found that the training session to be pretty informative. It's nice to gain a better understanding on the various topics we covered. The diagrams that were provided during the session helped to visually represent how processes flowed and related to one another. Although we flew through the installation, it would've been more beneficial if we were able to actually go along through the process on our own. That way if we were to run into actual issue with our own install we at least would have had the opportunity to maybe, with assistance, resolve any issues encountered. It was nice to see how much better the RPM install was compared to the old Pacman.
- Chris Pipes - I gained a firsthand perspective on problems someone could have while trying to install a CE and best practices to troubleshoot and correct those problems. Having tried to complete an install of the previous software, 1.2, and failing I was able to see the differences between Pacman and RPM/yum. I saw a great deal of improvement and an increase in the ease of use while installing the new software, 3.0. The introduction of the osg-system-profiler should prove to be a very useful tool for us in the future. It answers some very basic questions about the userís system, which will give OSG-Support a better understanding and a better starting point to provide front line support for the users. This will enable us to resolve more of the common problems quickly and easily. Our instructor, Alain, did a great job explaining the basics, what had changed from the previous version, going over the requirements and showing us the overall big picture. Walking us through the install of a CE was one of the most helpful things because he was able to answer some lingering questions that some of us have had for a while. Itís been a pleasure to have Alain here to train us and answer our questions. I will be please to find that our next training session is as helpful and fun as this has been.
- Scott Teige (Infrastructure) - The training was specifically structured to support staff and not toward infrastructure staff. This was by design and was absolutely appropriate. I was aware of the function and structure of most components discussed but had never seen them presented as a system and found that extremely useful. Infrastructure and Support interact on a daily basis but the training provided an environment where each could see operations from the others point of view. Exposed during the training was at least one case where Infrastructure, using services internally, communicated poorly with the user and VDT community and this alone (from the infrastructure position) was sufficient justification for the session. Finally, specific technical questions were discussed and solved and my take-away from that was that everyone involved became aware of where answers could be obtained who to get them.