Oracle SuperCluster Review

Zones rebooted randomly due to Clusterware bugs, and ZFS issues reduced performance


What is our primary use case?

Primary use case was a combination of a hardware refresh/consolidation of aging Sun M5000 equipment onto a half rack T5-8 SuperCluster. One of the core business drivers was the reduction of operational costs, while improving performance and (perceived) stability of the platforms.

What is most valuable?

The promise of consolidation is still a strong selling point, especially if the prospective client has applications that run on SPARC and has its data store in an Oracle Database.

What needs improvement?

There are many areas for improvement. For example, better guidance in terms of troubleshooting issues relating to ZFS, as well as better tools/diagnostics for monitoring that specific component, to better identify potential issues. Hardware monitoring via OEM 13.2 is not 100 percent, as Ops Center is still required in some instances.

For how long have I used the solution?

One to three years.

What do I think about the stability of the solution?

There was an issue with zones rebooting due to a CRS bug (Version 12.1.0.2 at the time) which took Oracle support quite a few weeks to eventually identify.

What do I think about the scalability of the solution?

While the general consensus was to maintain about 10 percent free space on ZFS file systems, it became clear that 30 percent free was actually required in order to maintain acceptable performance. Having to keep that headroom available impacted scalability to some extent.

How is customer service and technical support?

As with Oracle Support in general, it depends on the engineer picking up the call. Some issues were resolved quickly and with detailed information, while others often dragged and went around in circles.

Which solutions did we use previously?

Previous solution was M5000-based and the switch was from a consolidation/hardware refresh point of view.

How was the initial setup?

Initial setup had to be redone as there was some miscommunication and/or misunderstanding of requirements.

What other advice do I have?

Unfortunately, I have not had the best experience with Oracle’s SuperCluster. Zones were rebooting at random due to Clusterware bugs, there were various ZFS issues resulting in performance decreases and node evictions and, I’m sad to say, we did not receive the strongest vendor support in resolving the issues.

Do due diligence on sizing and whether or not the solution is, in fact, fit for purpose.

Disclosure: I am a real user, and this review is based on my own experience and opinions.
Add a Comment
Guest
Sign Up with Email