Re: Best fragmentation strategy

It depends. If your application is going after singleton rows, then there is no real advantage to spreading the load. If you are doing DSS (reading large chunks of data), then you want to take as much advantage of parallelism as possible - in other words spread the load over as many "spindles" as possible. I put that in quotes, because you don't have really good control over where the data is actually landing. A rule of thumb (for DSS) is 2 dbspaces per VCPU. j. On Aug 20, 2012, at 5:24 PM, Laurie Gustin wrote: > I have an opportunity to change the fragmentation strategy on some heavily used tables. Most records on these tables are only accessed on the day they are written. They are written, read, and updated only once or twice. Also keep in mind that we are on a SAN, so I have no control over where the data is physically stored. > > So my question is... is it better to have a different partition for each day of the month ( only one partition at a time is being heavily accessed) or to use round robin to theoretically 'spread out' the storage access across more of the disk. > > Thanks in advance for your opinions > > Laurie > _______________________________________________ > Informix-list mailing list > Informix-list@iiug.org > http://www.iiug.org/mailman/listinfo/informix-list