The question really is if your region server hosting the hot tail end of the region during sequential *writes* can take the load or not. If you find in the future that it cannot, manually splitting the regions is not going to fix the problem IMHO, since the tail end is always the one that is going to be hot in the case of sequential writes. At that point you will have to go with another strategy, like prefixing hashes.
On Sat, Nov 19, 2011 at 8:58 AM, Mark <[email protected]> wrote: > Also my reads will be far greater than writes so in this case is it fine > since I can manually split the region if there is a problem with reading? > > On 11/19/11 8:09 AM, Mark wrote: >> >> Right now we have 5 region servers and one of our tables is quite small >> (500k records) therefore it sits all in one region. This is happening >> because we are using a row key that corresponds to a FK in one of our MySQL >> databases so it's sequential in nature. I know the obvious fix for this >> would be to apply some sort of hash function on the rowkey before >> reading/writing but before I go down that route I wanted to know if I should >> bother. When does hot spotting start to become an issue? Would you say this >> is a cause for concern at this time? I should mentioned that at this time we >> are only writing to this table but we will soon be going live so it will >> start receiving get requests. >> >> Thanks for any suggestions. >
