A previous company that I worked for used DB2 Purescale which is basically HA DB2 with GPFS for the filesystem with crm for cluster management. On Sep 3, 2015 10:59 AM, "Zachary Giles" <[email protected]> wrote:
> Hello Everyone, > > Medium-time user of GPFS, MySQL, PostgreSQL, etc here.. Decent sized > system in production, hundreds of nodes, lots of tuning etc. Not a > newb. :) > Looking for opinions on running database engines backed by GPFS. > Has anyone run any backed by GPFS and what did you think about it? > > I realize there are tuning guides and guide-lines for running > different DBs on different file systems, but there seems to be a lack > of best-practices for doing so on GPFS. > > For example, usually you don't run DBs backed by NFS due to locking, > cacheing etc.. You can tune those out with sync, hard, etc, but, still > the best practice is to use a local file system. > As GPFS is hybrid, and used for many apps that do have hard > requirements such as Cinder block storage, science apps, etc, and has > proper byte-level locking.. it seems like it would be semi-equal to a > lock file system. > > Does anyone have any opinions, experiences, or recommendations for > running DBs backed by GPFS? > Also will accept horror stories, gotcha's, and "dont do it's". :) > > On that same note... > How about VMware? > Obviously I guess really the only way would be via NFS export.. which > cNFS was .. not the best at (my opinion). Maybe Protocol Servers are > better? Maybe also a "don't do it"? > > Thanks, > -Zach > > > -- > Zach Giles > [email protected] > _______________________________________________ > gpfsug-discuss mailing list > gpfsug-discuss at gpfsug.org > http://gpfsug.org/mailman/listinfo/gpfsug-discuss >
_______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at gpfsug.org http://gpfsug.org/mailman/listinfo/gpfsug-discuss
