So Sun continues their buying spree. First with CFS, and now Mysql. This is quite relevant to HPC for a variety of reasons.
Customers have been telling me that they are quite worried about Lustre as a result of the Sun acquisition. There is a lot of speculation about its (likely) limited future outside of Solaris.
Then there is ZFS, the sort of – kind of – open source file system that is sort of – kind of – better than anything else. Sort of – kind of . It still remains to be seen if it is really open source … I will leave the judgement of whether or not it is better for another day. FWIW: preferred file systems on Linux are xfs and jfs. ext3 is ok for small boot partitions if you like, and even every now and then, for the OS drive, such as on laptops. Xfs is and has been an industrial strength fs for a while, and at least the Irix version had some of the rudiments of volume management integrated in … though I never personally used them. You can see some of this in older manual pages on xfs.
The question is, Sun with its somewhat new, and uncertain record of OSS support and maintenance, its internal battle with Linux (that its customers are apparently deciding more in favor of Linux than Solaris, despite Solaris shipping default on every machine), what does this portend for Sun purchases? If our customers’ reaction to Lustre and CFS is any indicator, I expect to see Postgres usage on the sharp increase.
Of course, another important question is, what does Oracle think?
Mysql has been used by many HPC projects as a data repository. Many of these are open source. At least one of these, Rocks, depends critically upon it.
We are agnostic, our JackRabbits will run all manner of DB with great speed. I personally like SQLite3 for small tiny simple DBs, and Postgresql for more complex bits. The latter is quite nice, we are using it with DragonFly and other tools. FWIW: using the very nice DBIx::SimplePerl perl module (one of the better ones, wink wink) switching between DBs is very simple. We just did this with DragonFly, switching from SQLite3 to Postgres in a few minutes (mostly editing, and then fixing some add/updates to use NULLs where needed).
Not sure what the purchase portends, other than it could piss off a large Sun developer, and cause consternation within the OSS community. I hope I am wrong.