Our ASM Test Plan
In an upcoming blog, I'll discuss how important soft skills are to technicians. It is my personal opinion that no matter how strong of a technician you are, you must have strong inter-personal, communication and organizational skills to excel in our profession. Gone are the days when you could get by with just being a good "techie". I've been in the profession for twenty years now and continue to work on my soft skills. I just signed myself up for a class titled "Communicating with Diplomacy and Tact". I could always use a refresher on how to communicate in difficult and stressful situations.
You must communicate clearly in both verbal and written form. In addition, you also need to learn how to interact with your fellow employees in a professional manner. All jobs are stressful at times and disagreements in our profession are common. Being viewed as someone who "doesn't play well with others" will not help you advance your career. Trust me on this one. As I stated, I will be spending a blog, or two, on soft skills in the near future. But if you would like to start your soft skill education early, here's a link to an excellent article on the importance of soft skills for computer professionals.
The intent of our facilitated brainstorming session was to create an ASM test plan. The test plan will be decomposed into a set of detailed ASM test cases. Once we have our detailed test cases documented, I'll be sure to provide them to you in a future blog.
You'll notice that the term "PowerPath" shows up in the test plan. EMC is our storage vendor of choice. PowerPath is EMC's technology that provides multiple path I/O, automatic load balancing and path failover capabilities for servers connected to EMC Symmetrix and Clariion disk subsystems.
In addition, I have recently freed up enough time to begin reading the Oracle documentation on ASM. We have several large projects consuming a lot of my time, so I wasn't able to complete my personal education on ASM before Jim McQuade led our brainstorming session. I'm sure I'll rule out some of the bullets below (and add a few others) as I read the mountain of documentation I have amassed.
Administration
- ASM product installation
- ASM product configuration
- ASM instance startup
- ASM instance shutdown
- Database instance startup
- Database instance shutdown
- Operating system shutdown
- Operating system restart
- Migrating a database's files from a file system to ASM
- Migrating database's files from ASM to a file system
- Migrating additional databases to ASM without impacting existing ASM managed databases
- Detaching additional databases from ASM without impacting existing ASM managed databases
- Moving ASM managed databases and disk from one ASM instance to another ASM instance on a different hardware server
-
Configuration
changes while ASM is active
- Adding disk
- Removing disk
-
Upgrades
- ASM
- Managed database (Oracle release)
- Operating system
- EMC PowerPath
Monitoring
- I/O performance
- I/O distribution
- Hot spot monitoring - General hot spot monitoring output
- Hot sport monitoring - Does the execution of infrequent heavy workloads (SQLLDR, heavy batch updates) activate balancing or can it be set to not monitor?
- 10G EM reports/graphs
- Archive log processing
-
Performance reports,
alerts and notifications
- I/O thresholds
- Impact on segment statistics
- Impact on file I/O statistics
- I/O waits
- Statspack
- 10G EM I/O performance reporting
-
Availability
alerts and notifications
- ASM instance failure
- Unplanned disk removal
- Disk failure
- Disk path outage
- Single ASM instance process failure
- Bad database block
- SAN zoning changes
-
ASM provided
reports
- Performance statistics
- Object growth
- Environment details - Disk and file layout maps
and related information
Balancing
- Identify time when rebalance is actually performed
- Manually activating rebalancing
- Establishing time windows when rebalancing should not be performed
- Determine if rebalancing can be "throttled" to ensure application performance is not affected
- Rebalance override at disk level
- 10G R2's capability of supporting ASM/non-ASM managed disk in a single database instance
- Impact on server CPU, I/O and memory during balancing
Performance
- Validate quality of the balancing algorithm
-
ASM vs. non-ASM
performance
- Batch workloads
- Online workloads
- 1st time vs. ongoing (allow ASM to rebalance)
- Hot spot rebalancing
- Disk addition/disk removal
- Effects of wholesale disk additions
- Different types of disk (Symmetrix and Clariion)
Backup/Recovery
For those of you that read my blog on 10G
RMAN testing Part 1, you know I am highly concerned about the affect that
a new database release has on our ability to successfully backup and recover
our Oracle database environments. If there is one topic that will make me lie
awake nights, it is backup and recovery. Considering we have several multi-terabyte
warehouses, you can understand my concern. Well, you can now add ASM to my list
of concerns. I intend to find out EXACTLY what impact ASM has on our backup
and recovery strategies. I intend to execute the complete RMAN test plan contained
in the 10G
RMAN testing blog. A couple of the test cases that are related just to RMAN
may not be included, but the majority of the test plan will be executed and
the results documented. I'll let you know how it goes once the test plan is
complete.
Interoperability
- Access from existing Oracle administration tools (10G EM)
- EMC PowerPath path availability and load balancing
- Recovery from EMC PowerPath reconfiguration/rediscovery (e.g. devices change names)
- Netbackup
- Veritas Clustering Services
- Operating System
utilities (NMON, SAR, IOstat, etc)
Conclusion
I think our high-level test plan is a good start. I'm sure that I'll be tweaking the contents of the detailed test cases as I continue my education on ASM. If you would like us to add a test case for you, let me know. If I can test something for you, I will.
Thanks for reading,
Chris Foot