cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benedict (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7066) Simplify (and unify) cleanup of compaction leftovers
Date Wed, 12 Aug 2015 07:10:45 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14693042#comment-14693042
] 

Benedict commented on CASSANDRA-7066:
-------------------------------------

Looking good. I've pushed a few more minor stylistic suggestions [here|https://github.com/belliottsmith/cassandra/tree/7066-b-suggestions]

I think there's just one thing we need to address before commit: on a non-recoverable read
error of a transaction log, we can now fail to return any files for the directory. I think
we should make the Directories object have an option (that defaults to hard failure, via exception)
to determine behaviour in the event of such a failure. We should then make a conscious decision
about each client, as to what the desired behaviour should be. I suspect in many cases the
best thing we can do is to assume the txn log is not there after a few attempts, whilst logging
our error.

> Simplify (and unify) cleanup of compaction leftovers
> ----------------------------------------------------
>
>                 Key: CASSANDRA-7066
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7066
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Benedict
>            Assignee: Stefania
>            Priority: Minor
>              Labels: benedict-to-commit, compaction
>             Fix For: 3.0 alpha 1
>
>         Attachments: 7066.txt
>
>
> Currently we manage a list of in-progress compactions in a system table, which we use
to cleanup incomplete compactions when we're done. The problem with this is that 1) it's a
bit clunky (and leaves us in positions where we can unnecessarily cleanup completed files,
or conversely not cleanup files that have been superceded); and 2) it's only used for a regular
compaction - no other compaction types are guarded in the same way, so can result in duplication
if we fail before deleting the replacements.
> I'd like to see each sstable store in its metadata its direct ancestors, and on startup
we simply delete any sstables that occur in the union of all ancestor sets. This way as soon
as we finish writing we're capable of cleaning up any leftovers, so we never get duplication.
It's also much easier to reason about.



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

Mime
View raw message