cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carl Yeksigian (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (CASSANDRA-11575) Add out-of-process testing for CDC
Date Mon, 20 Jun 2016 21:02:58 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-11575?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Carl Yeksigian updated CASSANDRA-11575:
---------------------------------------
    Attachment: 11575.tgz

Given that there were a few changes to the API between when this was written and when CDC
was committed, I've updated the source to reflect that.

> Add out-of-process testing for CDC
> ----------------------------------
>
>                 Key: CASSANDRA-11575
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11575
>             Project: Cassandra
>          Issue Type: Sub-task
>          Components: Coordination, Local Write-Read Paths
>            Reporter: Carl Yeksigian
>            Assignee: Carl Yeksigian
>             Fix For: 3.x
>
>         Attachments: 11575.tgz, 11575.tgz
>
>
> There are currently no dtests for the new cdc feature. We should have some, at least
to ensure that the cdc files have a lifecycle that makes sense, and make sure that things
like a continually cleaning daemon and a lazy daemon have the properties we expect; for this,
we don't need to actually process the files, but make sure they fit the characteristics we
expect from them. A more complex daemon would need to be written in Java.
> I already hit a problem where if the cdc is over capacity, the cdc properly throws the
WTE, but it will not reset after the overflow directory is undersize again. It is supposed
to correct the size within 250ms and allow more writes.



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

Mime
View raw message