You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Alexey Scherbakov (Jira)" <ji...@apache.org> on 2021/08/10 10:15:00 UTC
[jira] [Updated] (IGNITE-15280) Git rid of sequential reads in raft
group
[ https://issues.apache.org/jira/browse/IGNITE-15280?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexey Scherbakov updated IGNITE-15280:
---------------------------------------
Description:
Due to some legacy issues all reads are implemented using [1] callback, efficiently making it sequential.
If a lease-based read (having lowest latency) model is used this is pointless, even for a readindex method reads can be concurrent.
We should refactor org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead for allowing concurrent batched reads.
Another task is to switch to lease reads as a default method.
[1] org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead
was:
Due to some legacy issues all reads are implemented using [1] callback,
efficiently making it sequential.
If a lease-based read (having lowest latency) model is used this is pointless, even for a readindex method reads can be concurrent.
We should refactor org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead for allowing concurrent batched reads.
Another task is to switch to lease reads as a default method.
[1] org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead
> Git rid of sequential reads in raft group
> -----------------------------------------
>
> Key: IGNITE-15280
> URL: https://issues.apache.org/jira/browse/IGNITE-15280
> Project: Ignite
> Issue Type: Improvement
> Affects Versions: 3.0.0-alpha2
> Reporter: Alexey Scherbakov
> Priority: Major
>
> Due to some legacy issues all reads are implemented using [1] callback, efficiently making it sequential.
> If a lease-based read (having lowest latency) model is used this is pointless, even for a readindex method reads can be concurrent.
> We should refactor org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead for allowing concurrent batched reads.
> Another task is to switch to lease reads as a default method.
> [1] org.apache.ignite.internal.table.distributed.raft.PartitionListener#onRead
--
This message was sent by Atlassian Jira
(v8.3.4#803005)