As a company that primarily supports MS SQL, Oracle and DB2 the promised reduced TOC of MySQL, and open source in-general, has caught the attention of management.
I am trying to evaluate MySQL and where it best fits (form and function) in our environment going forward. I don’t think “convert everything to open source” is necessarily a valid motto as some even at my company have espoused. However, I DO believe it has a place and want to define a boundry of best practice.
To that end, I am researching the internet and other sources to get a feel for enterprise readiness. I do not see a lot of unbiased facts regarding many topics I care about. My initial impressiosn are: good for read only with mild update and maybe has support challenges in complex environments. But these are based on impressions not fact. I’d like to get to the facts and am hoping they exist.
Of particular interest DOES ANYBODY HAVE REFERENCE TO ARTICLES, COMMENTARIES, CASE STUDIES OR PERSONAL FACTS-BASED EXPERIENCE to contribute to the following:
( I have a fair “features list comparison”, I am more interested in operational and/or technical information )
-
Of particular interest: Scalability of MySQL in a relatively heavy concurrent update environment. Particularly the Innodb engine. I have mild interest in ISAM but I believe it is best read-only due to table locking. Case studies? General experience? Break point? Most of our systems are a mixture of OLTP or DSS.
-
Backup/Recovery: The default mechanisms are wanting. For large db’s I assume the favorite options are file system snapshots using hardware, schleping out for the “hot backup” tool or replicating data to a remote site although. I don’t know if the later addresses user/app initiated data “oops” recovery scenarios very well and can be considered a complete backup. Any thoughts/research here would be interesting.
-
Tuning/Problem Solving: Commercial DB’s continue to improve instrumenting their source codes and providing tools to diagnose and capture performance data. How does MySQL stack up? How easy is it to diagnose a process ( not just a bad query ) Anybody pony up for MySQL Enterprise? - is it effective?
-
Support of Very Larger Databases ( or data sets ) and the ability to perform maintenance. Including things like backups, DDL changes, table redefinitions, index additions etc. I know the commercials db’s have spent considerable time investing in capabilities that allow some of these functions to be performed so as to reduce availability impact.
-
Backwards compability. Commercials DB’s have thier own challenges but they are usually announced and slowly depricated. I’ve seen some complaints on the WEB that version differences can cause code rewrites. Thoughts on product stability here from a feature perspective?
-
Independence from commercial vendors. I know there was hubbub about Sun close-sourcing some MySQL features but they backed off. Also, Oracle bought Innodb. I don’t know about Sun but I don’t trust Oracle. Any thoughts on independence for MySQL and effects on innovation, release schedules, feature kill etc?
Thanks to any and all that reply.