You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Thomas Steinmaurer (Jira)" <ji...@apache.org> on 2020/10/01 06:09:00 UTC
[jira] [Updated] (CASSANDRA-16153) Cassandra 4b2 - JVM options from
*.options not read/set
[ https://issues.apache.org/jira/browse/CASSANDRA-16153?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Thomas Steinmaurer updated CASSANDRA-16153:
-------------------------------------------
Description:
Trying out Cassandra 4 beta 2 with Java 8 (AdoptOpenJDK) in AWS.
{noformat}
NAME="Amazon Linux AMI"
VERSION="2018.03"
ID="amzn"
ID_LIKE="rhel fedora"
VERSION_ID="2018.03"
PRETTY_NAME="Amazon Linux AMI 2018.03"
ANSI_COLOR="0;33"
CPE_NAME="cpe:/o:amazon:linux:2018.03:ga"
HOME_URL="http://aws.amazon.com/amazon-linux-ami/"
{noformat}
It seems the Cassandra JVM results in using Parallel GC.
{noformat}
INFO [Service Thread] 2020-10-01 00:00:56,233 GCInspector.java:299 - PS Scavenge GC in 541ms. PS Old Gen: 5152844776 -> 5726724752;
WARN [Service Thread] 2020-10-01 00:00:56,234 GCInspector.java:297 - PS MarkSweep GC in 1969ms. PS Eden Space: 2111307776 -> 0; PS Old Gen: 5726724752 -> 2581334376; PS Survivor Space: 363850224 -> 0
{noformat}
Although {{jvm8-server.options}} is using CMS.
{noformat}
#################
# GC SETTINGS #
#################
### CMS Settings
-XX:+UseParNewGC
-XX:+UseConcMarkSweepGC
-XX:+CMSParallelRemarkEnabled
-XX:SurvivorRatio=8
-XX:MaxTenuringThreshold=1
-XX:CMSInitiatingOccupancyFraction=75
-XX:+UseCMSInitiatingOccupancyOnly
-XX:CMSWaitDuration=10000
-XX:+CMSParallelInitialMarkEnabled
-XX:+CMSEdenChunksRecordAlways
## some JVMs will fill up their heap when accessed via JMX, see CASSANDRA-6541
-XX:+CMSClassUnloadingEnabled
...
{noformat}
In Cassandra 3, default has been CMS.
So, possibly there is something wrong in reading/processing {{jvm8-server.options}}?
was:
Trying out Cassandra 4 beta 2 with Java 8 (AdoptOpenJDK) on Ubuntu 18.04 LTS. It seems the Cassandra JVM results in using Parallel GC. In Cassandra 3, default has been CMS.
Digging a bit further, it seems like the {{jvm8-server.options}} resp. {{jvm11-server.options}} files aren't used/processed in e.g. {{cassandra-env.sh}}.
E.g. in Cassandra 3.11, here we something like that in {{cassandra-env.sh}}.
{noformat}
# Read user-defined JVM options from jvm.options file
JVM_OPTS_FILE=$CASSANDRA_CONF/jvm.options
for opt in `grep "^-" $JVM_OPTS_FILE`
do
JVM_OPTS="$JVM_OPTS $opt"
done
{noformat}
Can't find something similar in {{cassandra-env.sh}} for Cassandra 4 beta2.
> Cassandra 4b2 - JVM options from *.options not read/set
> -------------------------------------------------------
>
> Key: CASSANDRA-16153
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16153
> Project: Cassandra
> Issue Type: Bug
> Components: Local/Scripts
> Reporter: Thomas Steinmaurer
> Priority: Normal
>
> Trying out Cassandra 4 beta 2 with Java 8 (AdoptOpenJDK) in AWS.
> {noformat}
> NAME="Amazon Linux AMI"
> VERSION="2018.03"
> ID="amzn"
> ID_LIKE="rhel fedora"
> VERSION_ID="2018.03"
> PRETTY_NAME="Amazon Linux AMI 2018.03"
> ANSI_COLOR="0;33"
> CPE_NAME="cpe:/o:amazon:linux:2018.03:ga"
> HOME_URL="http://aws.amazon.com/amazon-linux-ami/"
> {noformat}
> It seems the Cassandra JVM results in using Parallel GC.
> {noformat}
> INFO [Service Thread] 2020-10-01 00:00:56,233 GCInspector.java:299 - PS Scavenge GC in 541ms. PS Old Gen: 5152844776 -> 5726724752;
> WARN [Service Thread] 2020-10-01 00:00:56,234 GCInspector.java:297 - PS MarkSweep GC in 1969ms. PS Eden Space: 2111307776 -> 0; PS Old Gen: 5726724752 -> 2581334376; PS Survivor Space: 363850224 -> 0
> {noformat}
> Although {{jvm8-server.options}} is using CMS.
> {noformat}
> #################
> # GC SETTINGS #
> #################
> ### CMS Settings
> -XX:+UseParNewGC
> -XX:+UseConcMarkSweepGC
> -XX:+CMSParallelRemarkEnabled
> -XX:SurvivorRatio=8
> -XX:MaxTenuringThreshold=1
> -XX:CMSInitiatingOccupancyFraction=75
> -XX:+UseCMSInitiatingOccupancyOnly
> -XX:CMSWaitDuration=10000
> -XX:+CMSParallelInitialMarkEnabled
> -XX:+CMSEdenChunksRecordAlways
> ## some JVMs will fill up their heap when accessed via JMX, see CASSANDRA-6541
> -XX:+CMSClassUnloadingEnabled
> ...
> {noformat}
> In Cassandra 3, default has been CMS.
> So, possibly there is something wrong in reading/processing {{jvm8-server.options}}?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org