cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tyler Hobbs (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-8993) EffectiveIndexInterval calculation is incorrect
Date Thu, 19 Mar 2015 21:46:41 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-8993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14370187#comment-14370187
] 

Tyler Hobbs commented on CASSANDRA-8993:
----------------------------------------

For the fix, we'll need to do two things. First, stop overwriting summaries when the sstable
is not a 2.1 format.  Second, to recover from this bug, when we load sstables with the old
format, we need to check the summary against the actual index to verify that it's at full
sampling.  If not, trigger a rebuild of the summary.  Unfortunately, I think that will need
to be done every time until the sstables have been converted to the new format.  We could
avoid that by persisting a flag somewhere, but I don't think the additional complexity is
worth the savings in startup time.

I'll work on the patch for this.

> EffectiveIndexInterval calculation is incorrect
> -----------------------------------------------
>
>                 Key: CASSANDRA-8993
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8993
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>            Reporter: Benedict
>            Assignee: Benedict
>            Priority: Blocker
>             Fix For: 2.1.4
>
>         Attachments: 8993.txt
>
>
> I'm not familiar enough with the calculation itself to understand why this is happening,
but see discussion on CASSANDRA-8851 for the background. I've introduced a test case to look
for this during downsampling, but it seems to pass just fine, so it may be an artefact of
upgrading.
> The problem was, unfortunately, not manifesting directly because it would simply result
in a failed lookup. This was only exposed when early opening used firstKeyBeyond, which does
not use the effective interval, and provided the result to getPosition().
> I propose a simple fix that ensures a bug here cannot break correctness. Perhaps [~thobbs]
can follow up with an investigation as to how it actually went wrong?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message