cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ariel Weisberg (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (CASSANDRA-9109) Repair appears to have some of untested behaviors
Date Thu, 02 Apr 2015 22:59:52 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-9109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Ariel Weisberg updated CASSANDRA-9109:
--------------------------------------
    Description: 
There is AntiCompactionTest and a few single process unit tests, but they aren't very convincing.
Looking at the docs to nodetool it looks like there are a few different ways that repair could
operate that aren't explored. dtest wise there is repair_test and incremental_repair test
which do give some useful coverage, but don't do everything.

It's also the kind of thing you might like to see tested with some concurrent load to catch
interactions with everything else moving about, but a dtest may not be the right place to
do that.

  was:
There is AntiCompactionTest and a few single process unit tests, but they aren't very convincing.
Looking at the docs to nodetool it looks like there are a few different ways that repair could
operate that aren't explored.

It's also the kind of thing you might like to see tested with some concurrent load to catch
interactions with everything else moving about, but a dtest may not be the right place to
do that.


> Repair appears to have some of untested behaviors
> -------------------------------------------------
>
>                 Key: CASSANDRA-9109
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9109
>             Project: Cassandra
>          Issue Type: Test
>            Reporter: Ariel Weisberg
>
> There is AntiCompactionTest and a few single process unit tests, but they aren't very
convincing. Looking at the docs to nodetool it looks like there are a few different ways that
repair could operate that aren't explored. dtest wise there is repair_test and incremental_repair
test which do give some useful coverage, but don't do everything.
> It's also the kind of thing you might like to see tested with some concurrent load to
catch interactions with everything else moving about, but a dtest may not be the right place
to do that.



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

Mime
View raw message