[ 
https://issues.apache.org/jira/browse/IGNITE-23479?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Kirill Tkalenko updated IGNITE-23479:
-------------------------------------
    Description: 
We need to add a method to 
*org.apache.ignite.internal.metastorage.server.KeyValueStorage* that will 
perform steps 4.1 - 4.7 from IGNITE-23293.

Algorithm:
# If the storage is in the recovery state, then we set a new compaction 
revision (*KeyValueStorage#setCompactionRevision*) and complete the method.
2. Otherwise, adds a new task (A) to the WatcheEvent queue and complete the 
method.
3. Task (A) sets a new compaction revision 
(*KeyValueStorage#setCompactionRevision*) and adds a new task (B) to the 
compaction thread pool and completes.
4. Task (B) collects all read operations for a revision less than the new one 
and asynchronously begins to wait for their completion and starts a local 
compaction (*KeyValueStorage#compact*).
5. After the compaction is complete, notifies about its completion.

> Add an asynchronous compaction method to the KeyValueStorage
> ------------------------------------------------------------
>
>                 Key: IGNITE-23479
>                 URL: https://issues.apache.org/jira/browse/IGNITE-23479
>             Project: Ignite
>          Issue Type: Improvement
>            Reporter: Kirill Tkalenko
>            Assignee: Kirill Tkalenko
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.0
>
>
> We need to add a method to 
> *org.apache.ignite.internal.metastorage.server.KeyValueStorage* that will 
> perform steps 4.1 - 4.7 from IGNITE-23293.
> Algorithm:
> # If the storage is in the recovery state, then we set a new compaction 
> revision (*KeyValueStorage#setCompactionRevision*) and complete the method.
> 2. Otherwise, adds a new task (A) to the WatcheEvent queue and complete the 
> method.
> 3. Task (A) sets a new compaction revision 
> (*KeyValueStorage#setCompactionRevision*) and adds a new task (B) to the 
> compaction thread pool and completes.
> 4. Task (B) collects all read operations for a revision less than the new one 
> and asynchronously begins to wait for their completion and starts a local 
> compaction (*KeyValueStorage#compact*).
> 5. After the compaction is complete, notifies about its completion.



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

Reply via email to