Skip to content

Commit

Permalink
HBASE-28251 [SFT] Add description for specifying SFT impl during snap…
Browse files Browse the repository at this point in the history
…shot recovery (#5570)

Signed-off-by: Duo Zhang <zhangduo@apache.org>
Signed-off-by: Nihal Jain <nihaljain@apache.org>
Signed-off-by: Peter Somogyi <psomogyi@apache.org>
  • Loading branch information
wchevreuil committed Dec 11, 2023
1 parent 82a2ce1 commit 25c639f
Show file tree
Hide file tree
Showing 2 changed files with 23 additions and 4 deletions.
8 changes: 4 additions & 4 deletions src/main/asciidoc/_chapters/bulk_data_generator_tool.adoc
Original file line number Diff line number Diff line change
Expand Up @@ -18,8 +18,8 @@
* limitations under the License.
*/
////
== Bulk Data Generator Tool
[[BulkDataGeneratorTool]]
= Bulk Data Generator Tool
:doctype: book
:numbered:
:toc: left
Expand All @@ -29,7 +29,7 @@
This is a random data generator tool for HBase tables leveraging Hbase bulk load.
It can create pre-splited HBase table and the generated data is *uniformly distributed* to all the regions of the table.

=== How to Use
== Usage

[source]
----
Expand All @@ -53,7 +53,7 @@ hbase org.apache.hadoop.hbase.util.bulkdatagenerator.BulkDataGeneratorTool -t TE
hbase org.apache.hadoop.hbase.util.bulkdatagenerator.BulkDataGeneratorTool -t TEST_TABLE -mc 10 -r 100 -sc 10 -Dmapreduce.map.memory.mb=8192
----

=== How it Works
== Overview

==== Table Schema
Tool generates a HBase table with single column family, i.e. *cf* and 9 columns i.e.
Expand Down
19 changes: 19 additions & 0 deletions src/main/asciidoc/_chapters/store_file_tracking.adoc
Original file line number Diff line number Diff line change
Expand Up @@ -143,3 +143,22 @@ example, that would be as follows:
----
alter 'my-table', CONFIGURATION => {'hbase.store.file-tracker.impl' => 'FILE'}
----

### Specifying trackers during snapshot recovery

It's also possible to specify a given store file tracking implementation when recovering a snapshot
using the _CLONE_SFT_ option of _clone_snasphot_ command. This is useful when recovering old
snapshots, taken prior to a change in the global configuration, or if the snapshot has been
imported from a different cluster that had a different store file tracking setting.
Because snapshots preserve table and colum family descriptors, a simple restore would reload
the original configuration, requiring the additional steps described above to convert the
table/column family to the desired tracker implementation.
An example of how to use _clone_snapshot_ to specify the *FILE* tracker implementation
is shown below:

----
clone_snapshot 'snapshotName', 'namespace:tableName', {CLONE_SFT=>'FILE'}
----

NOTE: The option to specify the tracker during snapshot recovery is only available for the
_clone_snapshot_ command. The _restore_snapshot_ command does not support this parameter.

0 comments on commit 25c639f

Please sign in to comment.