Hive Metastore high availability

0
votes

Under cloudera, i'm trying to configure hive metastore to be highly available.

So i followed the tutorial bellow : http://www.cloudera.com/content/www/en-us/documentation/archive/cdh/4-x/4-7-1/CDH4-Installation-Guide/cdh4ig_hive_metastore_configure.html

My question is : do i need to add load balancers between hive metastore server and the replicated databases ?

Cause i think that we either need to give a list of URIs of the databases to hive metastore servers or to use a load balancer to switch to an other database in case of failure.

thanks in advance

1
hadoophiveclouderahigh-availabilitymetastore
I couldn't find anything HA-related by following the link you've provided. Perhaps you might want to look at cloudera.com/content/www/en-us/documentation/enterprise/latest/…. Take into account that metastore HA doesn't imply database HA. – facha
I did a miss copy/past, the link that you provided to me, is the one that i used to configure my Hive Meta Store HA, and the question is related to the communication with a HA database, so do i need to add loadbalancer between the HA Meta stores and HA databases ? – Mehdi TAZI

1 Answers

0
votes

I find a solution to my question :

actually the answer depends on the database, if the database manages HA and failover, there is no need to have loadbalancers , otherwise we need it in case of failure.

for the configuration part we can put as many uris as dbs in the hive-site.xml: 

<property>
  <name>javax.jdo.option.ConnectionURL</name>
  <value>jdbc:mysql://myhost/metastore,</value>
  <description>the URL of the MySQL database</description>
</property>