Hi Preeti,
3GB heap is too small for such setup. I would try 10-15GB, but that
depends on usage patterns. You have 50GB machine and assuming that you
do not run anything other than solr you have 30GB to spare on Solr and
still leave enough to OS to cache entire index.
The best way to do heap tuning is to monitor it. You can use standard
java tools, but if you prefer to get more insight into Solr/OS behavior,
you should use proper monitoring solution. There are several cloud
solutions that you can use even for free on small Solr setups. One such
product is our SPM <http://sematext.com/spm>.
HTH,
Emir
--
Monitoring * Alerting * Anomaly Detection * Centralized Log Management
Solr & Elasticsearch Support * http://sematext.com/
On 10.08.2016 12:41, preeti kumari wrote:
Hi,
I am using solr 5.2.1 in cloud mode with 3 shards on 3 different servers.
Each server is having 20 GB of data size . Total memory on each server is
around 50 GB.
Continuos updates and queries are being fired to solr.
We have been facing OOM issues due to heap issues.
args we use: giving 3 GB of max heap space on each solr server
java -server -Xss256k* -Xms3g -Xmx3g* -XX:NewRatio=3 -XX:SurvivorRatio=4
-XX:TargetSurvivorRatio=90 -XX:MaxTenuringThreshold=8
-XX:+UseConcMarkSweepGC -XX:+UseParNewGC -XX:ConcGCThreads=4
*-XX:ParallelGCThreads=4* -XX:+CMSScavengeBeforeRemark
-XX:PretenureSizeThreshold=64m -XX:+UseCMSInitiatingOccupancyOnly
-XX:CMSInitiatingOccupancyFraction=50 -XX:CMSMaxAbortablePrecleanTime=6000
-XX:+CMSParallelRemarkEnabled -XX:+ParallelRefProcEnabled
-XX:CMSFullGCsBeforeCompaction=1 -XX:CMSTriggerPermRatio=80 -verbose:gc
-XX:+PrintHeapAtGC -XX:+PrintGCDetails -XX:+PrintGCDateStamps
-XX:+PrintGCTimeStamps -XX:+PrintTenuringDistribution
-XX:+PrintGCApplicationStoppedTime ....
Please do let me know if these parameters values are enough for solr
running without OOM.
Let me know how can I fix these OOM issues.
Thanks
Preeti
