hbase-util

Utilites for making it easier to interact with hbase

Usage

Working with local hbase cluster

• lein repl

Working with a distributed hbase cluster

• mvn clean package
• Copy target/hbase-util-*.zip to target machine
• Unzip it on target machine
• kinit (if the cluster is secure)
• Run ./bin/hbase-util from within the extracted folder

bin/hbase-util expects..

• 'hbase' script to be on the classpath.
• HBASE_CONF_DIR and HADOOP_CONF_DIR env variables set.

Current set of features

Create hbase tables

> (doc create)
-------------------------
hbase-util.table.create/create
([f])
  Reads tables configuration from file 'f' and
creates the corresponding tables

> (create "conf/tables.yml")

Reset hbase tables

> (doc reset)
-------------------------
hbase-util.table.reset/reset
([f])
  Like (hbase shell's ) truncate, but doesn't delete/create
the tables, but instead, for each table..
  - Disables it
  - Deletes and re-creates all column familes
  - Enables the table again
This comes in handy if the logged-in user
has limited permissions. It also preserves
split information, unlike truncate.

> (reset "conf/tables.yml")

Verify hbase tables

> (doc verify)
-------------------------
hbase-util.table.verify/verify
([in out])
  Reads tables configuration from 'in' file, as yaml,
 and dumps any difference found in 'out' file, in yaml format

> (verify "conf/tables.yml" "/tmp/verfied.yml")

Configuration

Tables configuration is specified in yaml format see dev-tables.yml. This same yaml file serves both to 'create', 'reset' and 'verify' .

Splits for a table can be specified (in the input yaml conf file) either as

file: path/to/file

* ```
info: {negions: 16, algo: "HexStringSplit", first-row: "0", last-row "F"}

vals: ["0", "1", "2", "3"]

#### Split file
Table creation and validation routines assume that the non-printable characters in the split
file are hex-encoded. Note: this code honours lower-case hex digits like \xfe which is not
honoured by hbase currently