[
https://issues.apache.org/jira/browse/CASSANDRA-7066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14699744#comment-14699744
]
Benedict commented on CASSANDRA-7066:
-------------------------------------
bq. So the scenario is, we crash hard AND suffer xlog corruption so we don't
know which sstables are in-progress?
Right.
bq. (Is offline scrub xlog-aware? It probably should be.)
It is, but it hard fails on encountering a corrupted txn log; the operator can
then manually delete that log if they so desire (or move it aside, stash it,
whatever)
What about the sstables though? Right now we just leave them all there, but the
last "new" file may be partially written, which will end up crashing some read
queries. So the question is if we just fail and alert the user, or if we try to
establish that this is the case and stash those that are corrupted, or if we
just always move them aside.
> 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 that have been superceded); and 2) it's only used for a regular compaction -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)
