ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Eduard Shangareev (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (IGNITE-8114) Add fail recovery mechanism to tracking pages
Date Tue, 03 Apr 2018 13:53:00 GMT

    [ https://issues.apache.org/jira/browse/IGNITE-8114?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16424059#comment-16424059
] 

Eduard Shangareev commented on IGNITE-8114:
-------------------------------------------

[~dpavlov] 
I don't think so. 

Here I don't see any exception: BPlusTreePageMemoryImplTest. I guess, for some reason the
test took too much time. 
IgniteCacheEntryProcessorNodeJoinTest.testEntryProcessorNodeLeave - we had seen node from
other cluster. 
IgniteCacheServerNodeConcurrentStart.testConcurrentStart - PME hanged, no any issue with my
changes.

> Add fail recovery mechanism to tracking pages
> ---------------------------------------------
>
>                 Key: IGNITE-8114
>                 URL: https://issues.apache.org/jira/browse/IGNITE-8114
>             Project: Ignite
>          Issue Type: Improvement
>            Reporter: Eduard Shangareev
>            Assignee: Eduard Shangareev
>            Priority: Major
>             Fix For: 2.5
>
>
> Now we just assert that last saved tag in tracking page should be not greater than passed
one.
> But because of different issues which we don't understand yet, it could happen.
> So, I suggest to handle it by adding new "corruption" state to tracking state. 
> If tracking page is in a corrupted state than we would throw an exception on any querying
of the tracking page. Caller will have opportunity to catch the exception and recover page
state.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message