By related regions, do you mean query would typically involve both regions ?
Cheers On Jul 13, 2014, at 2:54 PM, gomes <[email protected]> wrote: > I have few tables, and I have pre partitioned the tables. I would like to > achieve reduced latency. I am just thinking of placing related regions on > the same region server/host. For example, table 1 contains 3 regions > [a,b,c], and table 2 contains 3 regions[x,y,z]. (a,x) contains the data for > some group of users, (b,y) contains the data for some group of users, and > (c,z) and so on. > > If I have two region servers, instead of placing [a,b,c] on one server, and > [x,y,z] on other server, is there an advantage of placing related regions > (a,x), (b,y), (c,z) among the region servers? > > Note: I also make sure the regions will not be divided further, or there > will be manual intervention needed to divide the regions further. > > > > -- > View this message in context: > http://apache-hbase.679495.n3.nabble.com/Is-there-an-advantage-of-placing-related-regions-on-the-same-region-server-host-tp4061296.html > Sent from the HBase User mailing list archive at Nabble.com.
