You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@phoenix.apache.org by anupama agarwal <an...@gmail.com> on 2016/06/14 11:23:02 UTC

Messed up index table due to merge_region command

Hi All,

I have interesting scenario and need your help for resolving this. I had
 created secondary index on my phoenix table with salt_buckets=20. Over
time due to wrong split policy and high load, my regions started splitting.
Number of regions for each index table increased from 20 to 60. Large
number of regions caused large number of store-files and my writes started
timing out to RPC timeout error.

I was suggested on Hbase-forum to merge regions manually using merge_region
command. I did that to reduce number of regions back. Merge was successful,
but I have an issue. My final index table has multiple entries for same
row. I am not able to understand why that happened. For example, this
result has been returned after using index table. ID field is unique and
should have only 1 value.


Now, in my main table, there is only 1 row for this id.


​

Can you please help me in this? How can fix my index data? Is the only
solution is to rebuild my index?
My HBase version is 1.0 and phoenix version 4.6.
​