[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15668175#comment-15668175 ] Shannon Carey commented on CASSANDRA-7254: -- Thanks. As it turns out, I think I had a similar issue as Jolokia where a rogue JVM agent created a non-daemon thread which prevented JVM termination despite presence of a JVM shutdown hook (those don't execute until all non-daemon threads have stopped). My initial theory regarding port usage seems to have been incorrect. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15667988#comment-15667988 ] Tyler Hobbs commented on CASSANDRA-7254: If you're using a version of Cassandra that includes CASSANDRA-10091, you can safely remove this check and still get a good error message (as confirmed in CASSANDRA-12074). Otherwise, this check is purely for cosmetic reasons, and you can remove it if it's causing problems. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15658039#comment-15658039 ] Shannon Carey commented on CASSANDRA-7254: -- This check causes Cassandra to freeze (block forever) if I use a JVM agent that tries to connect to the local JMX with `ManagementFactory.getPlatformMBeanServer()` (at least, I'm pretty sure that's why... it's definitely something to do with the agent.) It appears to only happen if we're using `com.sun.management.jmxremote.port` and not if we're only using `cassandra.jmx.local.port`. Can anyone explain why this happens? Is Cassandra doing something that causes other stuff to not be able to connect to the JMX port? Or is my agent's code not connecting properly? > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14191619#comment-14191619 ] viasyn commented on CASSANDRA-7254: --- Here is corresponding C* issue: https://github.com/rhuss/jolokia/issues/168 > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14191612#comment-14191612 ] viasyn commented on CASSANDRA-7254: --- Current workaround is really buggy. We are using Jolokia (JMX->HTTP bridge) with C* too. Here is example solution: https://github.com/causes/cassandra-nagios Jolokia prevents no-arg jvm from exit due to buggy daemon threads monitor so c* hangs on startup. The only acceptable workaround for now is to override "properties" system variable. This variable is undocumented and we afraid that it may disappear or change behavior. If you dont want to fix this workaround maybe you will add additional environment variable to customize launch_service argument? > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14164078#comment-14164078 ] Shaun Kalley commented on CASSANDRA-7254: - The 'grep -q' command in the patch is not Solaris 10 compatible. The common solution, I believe, is to direct the output of the grep command to /dev/null. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14122780#comment-14122780 ] Katsutoshi Nagaoka commented on CASSANDRA-7254: --- Thank you for this workaround. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14122521#comment-14122521 ] Chris McEniry commented on CASSANDRA-7254: -- Katsutoshi Nagaoka This is more that jolokia causes the preflight java run that was added in this issue to not exit and hangs up the startup script. The jvm-not-existing behavior happens even if you just try to startup java with only the -javaagent:jolokia.jar so I'm thinking it might be more for the jolokia crew to decide if that's intended behavior or not. Two thoughts on work around: 1.) (undo) You can out the preflight but then you'll see the NPEs that this was trying to avoid 2.) Instead of running jolokia as a commandline javaagent, have it attach after the start (http://www.jolokia.org/reference/html/agents.html#jvm-attach), but then you'll have to update the service start script or do some other manual start. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14117435#comment-14117435 ] Chris Lohfink commented on CASSANDRA-7254: -- [~na-ga] if you use {code} sudo netstat -anp | grep \* | grep LISTEN {code} you should be able to find whatever is using the port that either jmx or jolokia agent is using. Should probably switch to oracle JVM as well. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14117252#comment-14117252 ] Katsutoshi Nagaoka commented on CASSANDRA-7254: --- After CASSANDRA-7254, I can not start to cassandra-2.0.10 with jolokia agent. When I execute 'service cassandra start' command, startup process will be blocked by a grep. I think the cause of this problem that the next command is not completed. Is there a workaround something? my environment: * Cassandra 2.0.10 * Ubuntu 12.04.5 LTS * java version "1.7.0_65" * OpenJDK Runtime Environment (IcedTea 2.5.1) (7u65-2.5.1-4ubuntu1~0.12.04.2) * OpenJDK 64-Bit Server VM (build 24.65-b04, mixed mode) {code} root@ubuntu:~# grep jolokia /etc/cassandra/cassandra-env.sh # add the jolokia agent JVM_OPTS="$JVM_OPTS -javaagent:/usr/share/cassandra/lib/jolokia-jvm-1.2.0-agent.jar" root@ubuntu:~# "java" -cp /etc/cassandra:/usr/share/cassandra/lib/antlr-3.2.jar:/usr/share/cassandra/lib/commons-cli-1.1.jar:/usr/share/cassandra/lib/commons-codec-1.2.jar:/usr/share/cassandra/lib/commons-lang3-3.1.jar:/usr/share/cassandra/lib/compress-lzf-0.8.4.jar:/usr/share/cassandra/lib/concurrentlinkedhashmap-lru-1.3.jar:/usr/share/cassandra/lib/disruptor-3.0.1.jar:/usr/share/cassandra/lib/guava-15.0.jar:/usr/share/cassandra/lib/high-scale-lib-1.1.2.jar:/usr/share/cassandra/lib/jackson-core-asl-1.9.2.jar:/usr/share/cassandra/lib/jackson-mapper-asl-1.9.2.jar:/usr/share/cassandra/lib/jamm-0.2.5.jar:/usr/share/cassandra/lib/jbcrypt-0.3m.jar:/usr/share/cassandra/lib/jline-1.0.jar:/usr/share/cassandra/lib/jna-4.0.0.jar:/usr/share/cassandra/lib/jolokia-jvm-1.2.0-agent.jar:/usr/share/cassandra/lib/json-simple-1.1.jar:/usr/share/cassandra/lib/libthrift-0.9.1.jar:/usr/share/cassandra/lib/log4j-1.2.16.jar:/usr/share/cassandra/lib/lz4-1.2.0.jar:/usr/share/cassandra/lib/metrics-core-2.2.0.jar:/usr/share/cassandra/lib/netty-3.6.6.Final.jar:/usr/share/cassandra/lib/reporter-config-2.1.0.jar:/usr/share/cassandra/lib/servlet-api-2.5-20081211.jar:/usr/share/cassandra/lib/slf4j-api-1.7.2.jar:/usr/share/cassandra/lib/slf4j-log4j12-1.7.2.jar:/usr/share/cassandra/lib/snakeyaml-1.11.jar:/usr/share/cassandra/lib/snappy-java-1.0.5.jar:/usr/share/cassandra/lib/snaptree-0.1.jar:/usr/share/cassandra/lib/super-csv-2.1.0.jar:/usr/share/cassandra/lib/thrift-server-internal-only-0.3.3.jar:/usr/share/cassandra/apache-cassandra-2.0.10.jar:/usr/share/cassandra/apache-cassandra-thrift-2.0.10.jar:/usr/share/cassandra/apache-cassandra.jar:/usr/share/cassandra/stress.jar:/usr/share/java/jna.jar: -da -javaagent:/usr/share/cassandra/lib/jolokia-jvm-1.2.0-agent.jar -javaagent:/usr/share/cassandra/lib/jamm-0.2.5.jar -XX:+CMSClassUnloadingEnabled -XX:+UseThreadPriorities -XX:ThreadPriorityPolicy=42 -Xms3760M -Xmx3760M -Xmn400M -XX:+HeapDumpOnOutOfMemoryError -Xss256k -XX:StringTableSize=103 -XX:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:+CMSParallelRemarkEnabled -XX:SurvivorRatio=8 -XX:MaxTenuringThreshold=1 -XX:CMSInitiatingOccupancyFraction=75 -XX:+UseCMSInitiatingOccupancyOnly -XX:+UseTLAB -XX:+CMSParallelInitialMarkEnabled -XX:+CMSEdenChunksRecordAlways -XX:+UseCondCardMark -Djava.net.preferIPv4Stack=true -Dcom.sun.management.jmxremote.port=7199 -Dcom.sun.management.jmxremote.rmi.port=7199 -Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=false 2>&1 | grep -q 'Error: Exception thrown by the agent : java.lang.NullPointerException' ^C root@ubuntu:~# {code} > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1 rc3 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14053666#comment-14053666 ] Joshua McKenzie commented on CASSANDRA-7254: I slipped in logic to check for existing pid from pidfile when I rewrote the launch scripts so our only exposure on Windows is duplicate w/foreground mode. I'll open a ticket to replicate this logic on the Windows side. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1.0 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14051892#comment-14051892 ] T Jake Luciani commented on CASSANDRA-7254: --- +1 though not sure if we need the same for windows :) > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Brandon Williams >Priority: Minor > Fix For: 2.0.10, 2.1.0 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14045223#comment-14045223 ] Brandon Williams commented on CASSANDRA-7254: - Yeah, I would put it in 2.0. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0, 2.0.10 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14045212#comment-14045212 ] Jeremiah Jordan commented on CASSANDRA-7254: Since CASSANDRA-7087 went in to 2.0, I assume this will as well? > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > Attachments: 7254.txt > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14044164#comment-14044164 ] Chris Lohfink commented on CASSANDRA-7254: -- I did open a JVM ticket but hasnt gotten approved yet, I will update the ticket with info on it. I think the agent is worth exploring or having an alternative main class that can check the jmx port (provided as arg) to execute in cassandra-env.sh script. For CCM wont be a problem necessarily be a problem since the line that sets the rmi.port is in the cassandra-env.sh script which it doesnt use. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14044146#comment-14044146 ] Brandon Williams commented on CASSANDRA-7254: - I thought we could take a simple approach of looking for an existing Cassandra process, but even that won't work because we'll break CCM and people who run multiple instances per node. We can't really rely on the pidfile either because of foreground mode. Trapping stderr in the shell is difficult because at that point we're calling exec. We could have a small java program that tries to bind the ports and exits non-zero if it can't, then print out a message saying we couldn't bind before trying to start cassandra. > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14044103#comment-14044103 ] Tyler Hobbs commented on CASSANDRA-7254: Damn, none of those options are very good. bq. remove rmi port option until jvm fixed Most users won't be on the latest and greatest JVM, and JMX port madness has caused a lot of problems for users. bq. write a java app (platform independent, would have to work in many environments) that checks ports being open with appropariate and execute it in the bat/sh scripts to give a better warning This would need to use (parts of) the cassandra config and environment variables. I guess it's doable, just a pain. bq. custom java agent that would do kind of a "preflight check" on startup, giving some appropriate warnings if any ports are used etc We wouldn't have to set up a separate process for this, which is nice. My main concern is implementing all of the JMX features correctly in a custom agent, making sure it works across multiple java versions, and that future java versions don't break it. bq. ignore it until fixed (not ideal, but complexity for fixes might be worse then error) I guess Google will work reasonable well for users who hit this. I also played around with piping stderr through a process that could look for this particular string and output something more reasonable if found. It seems like this would be pretty hard to make portable across shells, even if we ignore windows. Somebody with more advanced shell-fu than I have might have a good idea, though. Did you go ahead and open a JVM ticket? > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14040372#comment-14040372 ] Chris Lohfink commented on CASSANDRA-7254: -- My thoughts: * remove rmi port option until jvm fixed * write a java app (platform independent, would have to work in many environments) that checks ports being open with appropariate and execute it in the bat/sh scripts to give a better warning * custom java agent that would do kind of a "preflight check" on startup, giving some appropriate warnings if any ports are used etc * ignore it until fixed (not ideal, but complexity for fixes might be worse then error) any feedback on a good approach? > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (CASSANDRA-7254) NPE on startup if another Cassandra instance is already running
[ https://issues.apache.org/jira/browse/CASSANDRA-7254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14037496#comment-14037496 ] Chris Lohfink commented on CASSANDRA-7254: -- Seems like it, two instances of: {code:title=Test.java|java} public class Test { public static void main(String[] args) throws Exception { Thread.sleep(10); } } {code} and jvm args {code}-Dcom.sun.management.jmxremote.port= -Dcom.sun.management.jmxremote.rmi.port= -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false{code} will cause that error. Ill open issue and see if theres something to do about it for a better error > NPE on startup if another Cassandra instance is already running > --- > > Key: CASSANDRA-7254 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7254 > Project: Cassandra > Issue Type: Bug > Components: Core >Reporter: Tyler Hobbs >Assignee: Chris Lohfink >Priority: Minor > Fix For: 2.1.0 > > > After CASSANDRA-7087, if you try to start cassandra while another instance is > already running, you'll see something like this: > {noformat} > $ bin/cassandra -f > Error: Exception thrown by the agent : java.lang.NullPointerException > {noformat} > This is probably a JVM bug, but we should confirm that, open a JVM ticket, > and see if we can give a more useful error message on the C* side. -- This message was sent by Atlassian JIRA (v6.2#6252)