You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Kirill Tkalenko (Jira)" <ji...@apache.org> on 2023/11/22 11:50:00 UTC
[jira] [Created] (IGNITE-20923) Possible optimization for sending row IDs during index building
Kirill Tkalenko created IGNITE-20923:
----------------------------------------
Summary: Possible optimization for sending row IDs during index building
Key: IGNITE-20923
URL: https://issues.apache.org/jira/browse/IGNITE-20923
Project: Ignite
Issue Type: Improvement
Reporter: Kirill Tkalenko
Now, when building an index, we send all row IDs from the row store, starting from the smallest to the end of the store, and we want to build indexes for all of them.
I think that we do not need to send all row IDs, but only those in which the version chains have versions less than or equal to the time the index was created. Newer ones will be sent in the normal operation of the cluster in the replication protocol or full rebalancing.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)