Thank for your answer. I don't know if I understand correctly. If the configurations are the same, does HBase share the connection?
________________________________ 보낸 사람: Ted Yu <yuzhih...@gmail.com> 보낸 날짜: 2017년 2월 4일 토요일 오후 9:04:52 받는 사람: user@hbase.apache.org 제목: Re: How should I pool Connection (HConnectionImplementation)? Did you observe some bottleneck w.r.t. using connection ? If you look at ConnectionManager.java, there is map from HConnectionKey to HConnectionImplementation. In hbase-client/src/main/java/org/apache/hadoop/hbase/client/HConnectionKey.java, you would see: * In essence, this class captures the properties in {@link Configuration} * that may be used in the process of establishing a connection. In light of * that, if any new such properties are introduced into the mix, they must be * added to the {@link HConnectionKey#properties} list. In short, there is already connection sharing underneath. FYI On Fri, Feb 3, 2017 at 10:45 PM, Kang Minwoo <minwoo.k...@outlook.com> wrote: > Good morning. > > > I'm using hbase-client 1.2.4 version. > > My client environment is multithreaded. > > I shared a connection, but I want to connection pooling to improve > performance. > > > Is there any good guide or best practice for connection pooling? > > Should I simply use Apache Commons Pool? > > > Let me know if I have something to consider. > > > Thank you all the time. > > > Outlook<http://aka.ms/weboutlook>에서 보냄 >