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 2022/11/17 07:01:00 UTC

[jira] [Updated] (IGNITE-18180) Implement partition destruction for RocsDB

     [ https://issues.apache.org/jira/browse/IGNITE-18180?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Kirill Tkalenko updated IGNITE-18180:
-------------------------------------
    Description: 
At the moment, the destruction of the partition does not work quite correctly, what are the problems:
# On *org.apache.ignite.internal.storage.rocksdb.RocksDbTableStorage#destroyPartition* call we don't destroy indexes;
# If the node falls after the call to destroy the partition and before flushing, then the partition will continue to exist, which may lead to negative consequences.

Also important:
# Write appropriate tests;
# On the destruction of the partition, we must not give the opportunity to read data from the partition and indexes and close all existing cursors for the partition and indexes.

  was:
At the moment, the destruction of the partition does not work quite correctly, what are the problems:
# On *org.apache.ignite.internal.storage.rocksdb.RocksDbTableStorage#destroyPartition* call we don't destroy indexes;
# If the node falls after the call to destroy the partition and before flushing, then the partition will continue to exist, which may lead to negative consequences.


> Implement partition destruction for RocsDB
> ------------------------------------------
>
>                 Key: IGNITE-18180
>                 URL: https://issues.apache.org/jira/browse/IGNITE-18180
>             Project: Ignite
>          Issue Type: Improvement
>            Reporter: Kirill Tkalenko
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.0.0-beta2
>
>
> At the moment, the destruction of the partition does not work quite correctly, what are the problems:
> # On *org.apache.ignite.internal.storage.rocksdb.RocksDbTableStorage#destroyPartition* call we don't destroy indexes;
> # If the node falls after the call to destroy the partition and before flushing, then the partition will continue to exist, which may lead to negative consequences.
> Also important:
> # Write appropriate tests;
> # On the destruction of the partition, we must not give the opportunity to read data from the partition and indexes and close all existing cursors for the partition and indexes.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)