Curious enough, when I leaved 5 of 10 nodes, everything goes well again.
No index missed.
But when I had 10 nodes, the read repair cannot help to find back the index.
I found RiakTS is still in very beginning stage, I nearly try to shift from
RiakKV + YZ to RiakTS.
Best regards,
Witeman
> On
Hello Fred,
riak-admin aae-status:
== Exchanges ==
Index Last (ago)All (ago)
---
0
Hello Witeman,
What you are seeing with your two queries is the result of two different
coverage plans, querying different parts of the cluster. Riak Search
translates coverage plans to Solr sharded queries, and will periodically change
the coverage plan, so as to more evenly distribute querie
Hi,
I am having a 10 nodes of RiakKV 2.2.0, and turn on Riak Search(Yokozuna).
Having about 3million records in one bucket with index, every record has about
1k size.
Then when it is triggered a Yokozuna query for one specific id, sometimes
return the record, sometimes return NOT FOUND, it i