Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase, mail # user - Format of -ROOT- and .META. tables


Copy link to this message
-
Format of -ROOT- and .META. tables
Vimal Jain 2014-03-18, 05:51
Hi,
Just while scanning ROOT and META tables on 2 node cluster , i found
following :-

scan '-ROOT-'

 .META.,,1                          column=info:regioninfo,
timestamp=1394103619228, value={NAME => '.META.,,1', STARTKEY => '', ENDKEY
                                    => '', ENCODED => 1028785192,}

 .META.,,1                          column=info:server,
timestamp=1394103628033,
value=ip-10-14-24-19.ap-southeast-1.compute.internal:60
                                    020

 .META.,,1                          column=info:serverstartcode,
timestamp=1394103628033, value=1394103612380
 .META.,,1                          column=info:v, timestamp=1394103619228,
value=\x00\x00

So there is one row with RowKey '.META.,,1' and 4 different columns for
this rowkey.
What these 4 columns signify , i understood little about cq 'server'.

I am assuming that , in my case , there is only one region for META so only
one row while scanning ROOT.What if i have multiple META regions there will
be one row for each region in ROOT ?
I tried splitting META from master UI at 60010 , but it did not split it
into 2.

Also scanning META gave similar output as above but this time row key
consisted of user table name.
Is there any guide in understanding structure of these system tables ?
Please help.
Thanks and Regards,
Vimal Jain