[ https://issues.apache.org/jira/browse/HBASE-10303?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-10303: -------------------------- Priority: Blocker (was: Major) Fix Version/s: 0.96.2 Making a blocker against 0.96.2.... If anyone has anything to add to Rural notes, it'd be appreciated ([~jmspaggi] Do we need to integrate your page into refguide?) > Have snappy support properly documented would be helpful to hadoop and hbase > users > ---------------------------------------------------------------------------------- > > Key: HBASE-10303 > URL: https://issues.apache.org/jira/browse/HBASE-10303 > Project: HBase > Issue Type: Task > Components: documentation > Reporter: Rural Hunter > Priority: Blocker > Fix For: 0.96.2 > > > The currentl document for configuring snappy > support(http://hbase.apache.org/book/snappy.compression.html) is not complete > and it's a bit obscure. IMO, there are several improvments can be made: > 1. Describe the relationship among hadoop,hbase,snappy. Is the snappy > actually needed by hadoop hdfs or hbase itself? That's to make clear if you > need to configure snappy support in hbase or hadoop. > 2. It didn't mention the default hadoop binary package is compiled without > snappy support and you need to compile it with snappy option manually. > Actually it didn't work with any native libs on 64 bits OS as the > libhadoop.so in the binary package is only for 32 bits OS(this of course is a > hadoop issue not hbase. but it's good to mention it.). > 3. In my experience, I actually need to install both snappy and > hadoop-snappy. So the doc lack of the steps to install hadoop-snappy. > 4. During my set up, I found difference where hadoop and hbase to pick up the > native lib files. hadoop picks those files in ./lib while hbase picks in > ./lib/[PLATFORM]. If it's correct, it can also be mentioned. -- This message was sent by Atlassian JIRA (v6.1.5#6160)