High swap usage on DT Server
What could be the cause? This doesn't happen open and when it happens, I restart the server... What else can I be doing?
free -mt
total used free shared buffers cached
Mem: 48294 48170 124 0 8 10062
-/+ buffers/cache: 38099 10195
Swap: 2047 1884 163
Total: 50342 50054 287
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
25309 dynatrac 19 0 15.5g 14g 16m S 8.7 31.0 33692:11 java
ps -ef | grep 25309
---------------------------------------------------
4250 3237 0 08:44 pts/1 00:00:00 grep 25309
546 25309 25283 99 Nov25 pts/0 23-09:32:18 /home/dynatrace/dynatrace-5.6.0/jre/bin/java -Xmx14080M -Xms14080M -XX:MaxPermSize=256M -XX:PermSize=256M -XX:+UseConcMarkSweepGC -XX:+DisableExplicitGC -XX:CMSInitiatingOccupancyFraction=60 -XX:CMSIncrementalSafetyFactor=40 -XX:+CMSClassUnloadingEnabled -Djava.endorsed.dirs=lib/endorsed -Djava.awt.headless=true -Dosgi.configuration.area=osgi -Dosgi.checkConfiguration=true -Dosgi.bundles=org.eclipse.equinox.common@2:start, org.eclipse.update.configurator@3:start, com.dynatrace.diagnostics.serverbootstrap.jar@4:start, org.eclipse.equinox.event@start, org.eclipse.equinox.ds@start -Declipse.ignoreApp=true -Dosgi.noShutdown=true -Djava.util.logging.manager=com.dynatrace.diagnostics.util.modern.LogManager -Dorg.osgi.framework.language=en -Dosgi.framework.extensions=com.dynatrace.diagnostics.bootstrap.hook -Dosgi.nl=en_US -Dosgi.nl.user=en_US -Duser.country=US -Duser.language=en -DproxyHost -DproxyPort -Dhttp.proxyUser -Dhttp.proxyPassword -server -XX:+UseCompressedOops -XX:+OptimizeStringConcat -XX:+UseStringCache -XX:-UseBiasedLocking -Dcom.dynatrace.diagnostics.memory.corridorUpperLimit=90 -Dcom.dynatrace.diagnostics.memory.corridorReductionArea=85 -Dcom.dynatrace.diagnostics.memory.corridorNeutralArea=75 -Dcom.dynatrace.diagnostics.server.realtime.maxWorkerThreads=8 -Dcom.dynatrace.diagnostics.repository.cleanup.measurementhigh.delete.groupsize=20 -Dcom.dynatrace.diagnostics.repository.debugcleanupjob=true -Dcom.dynatrace.diagnostics.repository.cleanup.baseline.delete.groupsize=10 -Dcom.dynatrace.diagnostics.completenessTimeout=360 -Dcom.dynatrace.diagnostics.WatchDogTimeout=180 -agentpath:/home/dynatrace/dynatrace-5.6.0/server/lib/libdtserverhelperagent.so -Dosgi.os=linux -Dosgi.ws=gtk -Dosgi.arch=x86 -Dcom.dynatrace.diagnostics.WatchDogPort=50015 -jar /home/dynatrace/dynatrace-5.6.0/server/plugins/org.eclipse.equinox.launcher_1.2.0.v20110502.jar -name dtserver -showsplash -launcher /home/dynatrace/dynatrace-5.6.0/dtserver
Jibi
Server memory usage has not changed as per the self monitoring dash. There are few collectors on the box. They also don't show anything out of the norm.
Thanks to a colleague, got some details.... pasting below for all future such questions..
---------------------------------------------------------------------------------------------
We are not worried about a server using swap when there is plenty of memory available in the system cache/buffers. Applications do not cause swapping and it is pointless to point out which apps are using the most swap. Their use of swap is a result of something else requesting real memory ( most likely someone using "vim" or "less" looking at a large log files, file copying, a deployment, etc) so the kernel decided to use swap instead of giving up its buffers or cache.
Swapping is controlled by the kernel /proc/sys/vm/swappiness parameter.
swappiness
A value from 0 to 100 which controls the degree to which the system swaps. A high value prioritizes
system performance, aggressively swapping processes out of physical memory when they are not
active. A low value prioritizes interactivity and avoids swapping processes out of physical memory for
as long as possible, which decreases response latency. The default value is 60.
If there is > 512mb to 1gb of free memory in buffers/cache then swapping is not a problem.
