Hi
I am trying out the elasticsearch 1.0's snapshot and restore feature.
My doubt is when the restore operation is performed will the missing
shards, those available when snapshot was taken, restored?
When i tried this the missing shards were not restored. So does the restore
assumes all the shards are available and just restore the records?
You can restore an index if it doesn't currently exist. If it currently
exists, it must have the same shard count as when you ran the snapshot, and
in this case, restore will replace your data in the index to the state of
the specific snapshot that you choose to restore.
Thank you Binh Ly.
But i am correct to say a restore process should also restore missing
shards. Because in a case where if some shards are missing and restore
didn't help?
On Friday, 7 March 2014 20:58:15 UTC+5:30, Binh Ly wrote:
You can restore an index if it doesn't currently exist. If it currently
exists, it must have the same shard count as when you ran the snapshot, and
in this case, restore will replace your data in the index to the state of
the specific snapshot that you choose to restore.
As long as you have the same shard count as the backup and all the shards
are allocated, the restore will work properly. If you have a missing shard
(i.e. you deleted the underlying fs directory) and it is not currently
allocated, you can allocate it first using cluster reroute:
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.