Saturday, July 10, 2010

Oracle RAC For PeopleSoft - Part 1

So it's been several months since my last blog entry.  Things have been crazy at work, at home, and with IOUG and unfortunately, blogging is somewhat low on the priority list.  So, what have I been up to?  Well, the home life  would be enough for its own blog with two kids that are very involved in sports.  That is another story though.  Work has been very interesting with a couple projects.  The one I am in the middle of now is implementing Oracle Real Application Clusters (RAC) to support our PeopleSoft database.

I have been a DBA for about 13 years.  It is almost crazy to say that, but okay, I accept it.  I am on my second time around supporting a PeopleSoft environment.  I work for a company that is made up of many divisions and is part of a larger organization made up of several companies.  We have been consolidating financial systems into a PeopleSoft 9.0 Financials system from our divisions, and recently, from one of our sister companies.  Not only is it a challenge to integrate business processes between divisions and companies, but technically, we have been challenged to find a solution that will grow as we continue to consolidate.  The biggest part of that challenge is that we are not sure, from an IT perspective, when we will be consolidating.  We do not want to oversize our architecture now for where we are, but do not want to  undersize for future growth.

Performance of the PeopleSoft database has been questionable since I started at this company two years ago.  We have done quite a bit with indexes, changing some business processes, and actually changing some of the application processes.  A couple big wins involved converting PeopleSoft "temp" tables to Oracle temporary tables.  Even with all the tuning successes, adding many users through growth and consolidation has made it very apparent that we have outgrown our database environment.  This is really not bad considering the environment was implemented over three years ago with limited information.

Given that we had exhausted most of our tuning options and we knew we were bringing on new projects and users, we decided that new hardware was required.  The decision came down to a new, dedicated database server or Real Application Clusters.  The company is still working to consolidate IT operations, so we are not sure what divisions or how many users we will be supporting over the next three to five years.  Since we did not want to suggest spending several hundred thousand dollars on a new server this year, only to have to do it again in a year if we consolidated with another large division, RAC was the logical choice.

This entry has been in the works for several weeks and is getting long enough.  In the coming entries I will talk about how we put together a proof of concept, designed the architecture, built the environment and how the implementation is going.  Stay tuned.  I will make sure the follow-up entries are faster in coming than this one...

No comments:

Post a Comment