[
https://issues.apache.org/jira/browse/CASSANDRA-7066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644623#comment-14644623
]
Benedict commented on CASSANDRA-7066:
-------------------------------------
bq. Is that SOL as in "Now we can't tell which is new so we have to keep both
and do redundant compaction" or as in "Now the note can't start up?"
Well, SOL in that we have no way of guaranteeing C* will be able to notice this
has happened and do anything to avoid breaking things. If we see just one of
the two files present, under non-fs failure that means we were part way through
cleaning up, and need to finish deleting its contents. So if we lose the "new"
file, say, then we will delete the old files on startup none-the-wiser. The
result being a partial replacement of the sstables (perhaps with nothing at
all). There's really nothing you can do to guarantee anything in the face of a
filesystem corruption.
If the operator _knows this has happened_ before starting up they can delete
all the log files, sure, and then we'll keep everything.
> 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 that have been superceded); and 2) it's only used for a regular compaction -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)
