cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robert Stupp (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-11552) Reduce amount of logging calls from ColumnFamilyStore.selectAndReference
Date Sat, 07 May 2016 07:52:12 GMT


Robert Stupp commented on CASSANDRA-11552:

At least the situation I've seen is that someone ran offline-scrub while the node was up.
This caused the loop to spin forever and log as fast as it can. Due to the vast amount of
never ending messages, people tend assume a bug in the code.
I'm not objecting that the operator is alerted, but having 200MB log messages per minute repeating
the same message over and over again feels wrong.
I completely agree that this message indicates a very severe condition.
Frankly, that "offline scrub" situation and a "normal" slowdown are completely different situations.
The first will never recover whereas the other is kind of expected.

> Reduce amount of logging calls from ColumnFamilyStore.selectAndReference
> ------------------------------------------------------------------------
>                 Key: CASSANDRA-11552
>                 URL:
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Robert Stupp
>            Assignee: Robert Stupp
>             Fix For: 2.1.15, 2.2.7, 3.7, 3.0.7
> {{org.apache.cassandra.db.ColumnFamilyStore#selectAndReference}} logs two messages at
_info_ level "as fast as it can" if it waits for more than 100ms.
> The following code is executed in a while-true fashion in this case:
> {code}
>       "Spinning trying to capture released readers {}", released);
>       "Spinning trying to capture all readers {}", view.sstables);
> {code}

This message was sent by Atlassian JIRA

View raw message