cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marcus Eriksson (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-8429) Some keys unreadable during compaction
Date Fri, 19 Dec 2014 12:35:13 GMT


Marcus Eriksson commented on CASSANDRA-8429:

LGTM, few nits;

* Could you explain what is going on in CompressionMetadata#resetAndTruncate? (I guess it
is to handle the refcounting, making sure we are the only one holding a ref to the memory,
a comment explaining it would be nice)
* Using a queue or a stack for popping stuff out of finishedEarly instead of the destroy(..)-thing
would make it more obvious what is going on, without having to read the code in destroy(...)

> Some keys unreadable during compaction
> --------------------------------------
>                 Key: CASSANDRA-8429
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>         Environment: Ubuntu 14.04
>            Reporter: Ariel Weisberg
>            Assignee: Marcus Eriksson
>             Fix For: 2.1.3
>         Attachments: cluster.conf,
> Starts as part of merge commit 25be46497a8df46f05ffa102bc645bfd684ea48a
> Stress will say that a key wasn't validated because it isn't returned even though it's
loaded. The key will eventually appear and can be queried using cqlsh.
> Reproduce with
> #!/bin/sh
> ROWCOUNT=10000000
> SCHEMA='-col n=fixed(1) -schema compaction(strategy=LeveledCompactionStrategy) compression=LZ4Compressor'
> ./cassandra-stress write n=$ROWCOUNT -node xh61 -pop seq=1..$ROWCOUNT no-wrap -rate threads=25
> ./cassandra-stress mixed "ratio(read=2)" n=100000000 -node xh61 -pop "dist=extreme(1..$ROWCOUNT,0.6)"
-rate threads=25 $SCHEMA

This message was sent by Atlassian JIRA

View raw message