[
https://issues.apache.org/jira/browse/IGNITE-1510?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ivan Veselovsky updated IGNITE-1510:
------------------------------------
Description:
Currently we have two methods to remove something from IGFS:
1) remove - performs soft delete for PRIMARY mode and hard delete for others
2) format - delete of all IGFS data without touching seocndary file system,
which can be either soft or hard depending on some very coutner-intuitive
conditions.
I think we should do the following:
1) remove operation stays as is.
2) format method is deprecated and just falls-back to a new method
"clear(ROOT)".
3) "clear" operation is semantically identical to cache clear: remove in-memory
data, do not touch persistence layer. Essentially it just moves a tree into the
trash just like remove does. But also this operation will offer sync and async
modes. In sync mode operation exits when all in-memory data is really removed
even from trash.
was:
Currently we have two methods to remove something from IGFS:
1) remove - performs soft delete for PRIMARY mode and hard delete for others
2) format - delete of all IGFS data without touching seocndary file system,
which can be either soft or hard depending on some very coutner-intuitive
conditions.
I think we should do the following:
1) remove operation stays as is.
2) format method is deprecated and just falls-back to a new method
"clear(ROOT)".
3) "clear" operation is semantically identical to cache clear: remove in-memory
data, do not touch persistence layer. Essentially it just moves a tree into the
trash just like remove does. But also this operation will offer sync and async
modes. In sync mode operation exits when all data in-memory data is really
removed even from trash.
> Weird format() and remove() semantics.
> --------------------------------------
>
> Key: IGNITE-1510
> URL: https://issues.apache.org/jira/browse/IGNITE-1510
> Project: Ignite
> Issue Type: Task
> Components: hadoop
> Affects Versions: 1.1.4
> Reporter: Vladimir Ozerov
> Assignee: Ivan Veselovsky
> Priority: Critical
> Fix For: 1.5
>
>
> Currently we have two methods to remove something from IGFS:
> 1) remove - performs soft delete for PRIMARY mode and hard delete for others
> 2) format - delete of all IGFS data without touching seocndary file system,
> which can be either soft or hard depending on some very coutner-intuitive
> conditions.
> I think we should do the following:
> 1) remove operation stays as is.
> 2) format method is deprecated and just falls-back to a new method
> "clear(ROOT)".
> 3) "clear" operation is semantically identical to cache clear: remove
> in-memory data, do not touch persistence layer. Essentially it just moves a
> tree into the trash just like remove does. But also this operation will offer
> sync and async modes. In sync mode operation exits when all in-memory data is
> really removed even from trash.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
