# Optimize elasticsearch / JVM

**URL:** <https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874>\
**Category:** Elasticsearch\
**Created:** [January 28, 2015, 1:09pm UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874 "2015-01-28T13:09:21Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![oto\_iashvili](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@oto\_iashvili](https://discuss.elastic.co/u/oto_iashvili)\
**Post date:** [January 28, 2015, 1:09pm UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/1 "2015-01-28T13:09:21Z")

</div>

Hi

I have a website for classified. For this I'm using elasticsearch, postgres  
and rails on a same ubuntu 14.04 dedicated server, with 256go of RAM and 20  
cores, 40 threads .

I have 10 indexes on elasticsearch, each have default numbers of shardes  
(5). They have between 1000 and 400 000 classifieds dependings on which  
index.  
approximatly 5000 requests per minute, 2/3 making an elasticsearch request.

according to htop, jvm is using around 500% of CPU  
I try different options, I reduce number of shardes per index, I also try  
to change JAVA\_OPTS of followed

```
#JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
#JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"

#JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
#JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"

JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"

```

but it doesnt seems to change anything.

so to questions :

- when you change any setting on elasticsearch, and then restart, should  
the improvement (if any) be visible immediatly or can it arrive a bit later  
thanks to cache or any thing else ?
- can any one help me to find good configuration for JVM / elasticsearch so  
it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/369fdad8-cc02-415c-b4e9-e93135e58b59%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/369fdad8-cc02-415c-b4e9-e93135e58b59%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![Jilles\_van\_Gurp](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/jilles_van_gurp/32/879_2.png) [@Jilles\_van\_Gurp](https://discuss.elastic.co/u/Jilles_van_Gurp)\
**Post date:** [January 28, 2015, 1:36pm UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/2 "2015-01-28T13:36:29Z")

</div>

How much heap are you giving to ES? With this many requests, if your setup  
is not falling over it is probably not garbage collect related because that  
would result in very noticable delays/unavailability of es. 32GB should be  
a good value given how much memory you have. Also, you probably want to use  
doc\_values in your mapping so that you can utilize the os file cache and  
move some of the memory pressure on the heap. You seem to have plenty of  
ram, so your entire dataset should easily fit in RAM.

Also, don't use G1 for elasticsearch. There are known issues with that  
particular garbage collector in combination with lucene. CMS is the best  
option for ES.

500% of 20 cores doesn't sound that bad; you'd max them out at 4000%.  
Still, it would be nice to know what it is doing. In any case, you might  
want to try out marvel to find out where your setup is bottlenecked. Also,  
you might want to consider scaling horizontally instead of vertically. Many  
smaller servers can be nicer than one big one.

On Wednesday, January 28, 2015 at 2:09:22 PM UTC+1, Oto Iashvili wrote:

> Hi
> 
> I have a website for classified. For this I'm using elasticsearch,  
> postgres and rails on a same ubuntu 14.04 dedicated server, with 256go of  
> RAM and 20 cores, 40 threads .
> 
> I have 10 indexes on elasticsearch, each have default numbers of shardes  
> (5). They have between 1000 and 400 000 classifieds dependings on which  
> index.  
> approximatly 5000 requests per minute, 2/3 making an elasticsearch request.
> 
> according to htop, jvm is using around 500% of CPU  
> I try different options, I reduce number of shardes per index, I also try  
> to change JAVA\_OPTS of followed
> 
> ```
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
> 
> #JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
> 
> JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
> 
> ```
> 
> but it doesnt seems to change anything.
> 
> so to questions :
> 
> - when you change any setting on elasticsearch, and then restart, should  
> the improvement (if any) be visible immediatly or can it arrive a bit later  
> thanks to cache or any thing else ?
> - can any one help me to find good configuration for JVM / elasticsearch  
> so it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/49d08b16-276c-4639-8c1b-26cb5d74530f%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/49d08b16-276c-4639-8c1b-26cb5d74530f%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![oto\_iashvili](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@oto\_iashvili](https://discuss.elastic.co/u/oto_iashvili)\
**Post date:** [January 28, 2015, 10:41pm UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/3 "2015-01-28T22:41:16Z")

</div>

Hi,

thanks a lot for answer,  
Ive tried several value for heap, between 26 and 32, but I didnt see any  
difference.

I remove G1 and put back default parameter. But still same pb.  
I was saying around 500%, but it is just average, i goes sometimes up to  
2000%

I was also thinking to take several server, but right now it is not  
possible.

I was using a smaller server just before, with 96go of ram and I was  
working better. I trie to put same parameters as before, but it is not much  
slower.

On Wednesday, January 28, 2015 at 2:36:29 PM UTC+1, Jilles van Gurp wrote:

> How much heap are you giving to ES? With this many requests, if your setup  
> is not falling over it is probably not garbage collect related because that  
> would result in very noticable delays/unavailability of es. 32GB should be  
> a good value given how much memory you have. Also, you probably want to use  
> doc\_values in your mapping so that you can utilize the os file cache and  
> move some of the memory pressure on the heap. You seem to have plenty of  
> ram, so your entire dataset should easily fit in RAM.
> 
> Also, don't use G1 for elasticsearch. There are known issues with that  
> particular garbage collector in combination with lucene. CMS is the best  
> option for ES.
> 
> 500% of 20 cores doesn't sound that bad; you'd max them out at 4000%.  
> Still, it would be nice to know what it is doing. In any case, you might  
> want to try out marvel to find out where your setup is bottlenecked. Also,  
> you might want to consider scaling horizontally instead of vertically. Many  
> smaller servers can be nicer than one big one.
> 
> On Wednesday, January 28, 2015 at 2:09:22 PM UTC+1, Oto Iashvili wrote:
> 
> > Hi
> > 
> > I have a website for classified. For this I'm using elasticsearch,  
> > postgres and rails on a same ubuntu 14.04 dedicated server, with 256go of  
> > RAM and 20 cores, 40 threads .
> > 
> > I have 10 indexes on elasticsearch, each have default numbers of shardes  
> > (5). They have between 1000 and 400 000 classifieds dependings on which  
> > index.  
> > approximatly 5000 requests per minute, 2/3 making an elasticsearch  
> > request.
> > 
> > according to htop, jvm is using around 500% of CPU  
> > I try different options, I reduce number of shardes per index, I also try  
> > to change JAVA\_OPTS of followed
> > 
> > ```
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
> > 
> > #JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
> > 
> > JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
> > 
> > ```
> > 
> > but it doesnt seems to change anything.
> > 
> > so to questions :
> > 
> > - when you change any setting on elasticsearch, and then restart, should  
> > the improvement (if any) be visible immediatly or can it arrive a bit later  
> > thanks to cache or any thing else ?
> > - can any one help me to find good configuration for JVM / elasticsearch  
> > so it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/da7e20bf-c8b9-43ed-a95e-49ec32b3660c%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/da7e20bf-c8b9-43ed-a95e-49ec32b3660c%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![oto\_iashvili](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@oto\_iashvili](https://discuss.elastic.co/u/oto_iashvili)\
**Post date:** [January 29, 2015, 1:41am UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/4 "2015-01-29T01:41:00Z")

</div>

if it can help :

\_nodes/hot\_threads?pretty=1

::: [Sepulchre][tcr0m37nRAS9appE1pvWLA][sd-70975][inet[/x.x.x.x:9300]]

20.6% (102.7ms out of 500ms) cpu usage by thread  
'elasticsearch[Sepulchre][search][T#78]'  
8/10 snapshots sharing following 10 elements  
sun.misc.Unsafe.park(Native Method)  
java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)  
java.util.concurrent.LinkedTransferQueue.awaitMatch(  
LinkedTransferQueue.java:735)  
java.util.concurrent.LinkedTransferQueue.xfer(LinkedTransferQueue.  
java:644)  
java.util.concurrent.LinkedTransferQueue.take(LinkedTransferQueue.  
java:1137)  
org.elasticsearch.common.util.concurrent.SizeBlockingQueue.take(  
SizeBlockingQueue.java:162)  
java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.  
java:1068)  
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.  
java:1130)  
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor  
.java:615)  
java.lang.Thread.run(Thread.java:745)  
2/10 snapshots sharing following 15 elements  
org.apache.lucene.search.TopFieldCollector.populateResults(  
TopFieldCollector.java:1203)  
org.apache.lucene.search.TopDocsCollector.topDocs(TopDocsCollector.  
java:156)  
org.apache.lucene.search.TopDocsCollector.topDocs(TopDocsCollector.  
java:93)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:582)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:533)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:510)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:345)  
org.elasticsearch.search.query.QueryPhase.execute(QueryPhase.java:150  
)  
org.elasticsearch.search.SearchService.executeQueryPhase(  
SearchService.java:275)  
org.elasticsearch.search.action.SearchServiceTransportAction$5.call(  
SearchServiceTransportAction.java:231)  
org.elasticsearch.search.action.SearchServiceTransportAction$5.call(  
SearchServiceTransportAction.java:228)  
org.elasticsearch.search.action.SearchServiceTransportAction$23.run(  
SearchServiceTransportAction.java:559)  
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.  
java:1145)  
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor  
.java:615)  
java.lang.Thread.run(Thread.java:745)

20.2% (101ms out of 500ms) cpu usage by thread  
'elasticsearch[Sepulchre][search][T#4]'  
10/10 snapshots sharing following 10 elements  
sun.misc.Unsafe.park(Native Method)  
java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)  
java.util.concurrent.LinkedTransferQueue.awaitMatch(  
LinkedTransferQueue.java:735)  
java.util.concurrent.LinkedTransferQueue.xfer(LinkedTransferQueue.  
java:644)  
java.util.concurrent.LinkedTransferQueue.take(LinkedTransferQueue.  
java:1137)  
org.elasticsearch.common.util.concurrent.SizeBlockingQueue.take(  
SizeBlockingQueue.java:162)  
java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.  
java:1068)  
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.  
java:1130)  
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor  
.java:615)  
java.lang.Thread.run(Thread.java:745)

19.3% (96.4ms out of 500ms) cpu usage by thread  
'elasticsearch[Sepulchre][search][T#42]'  
9/10 snapshots sharing following 10 elements  
sun.misc.Unsafe.park(Native Method)  
java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)  
java.util.concurrent.LinkedTransferQueue.awaitMatch(  
LinkedTransferQueue.java:735)  
java.util.concurrent.LinkedTransferQueue.xfer(LinkedTransferQueue.  
java:644)  
java.util.concurrent.LinkedTransferQueue.take(LinkedTransferQueue.  
java:1137)  
org.elasticsearch.common.util.concurrent.SizeBlockingQueue.take(  
SizeBlockingQueue.java:162)  
java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.  
java:1068)  
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.  
java:1130)  
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor  
.java:615)  
java.lang.Thread.run(Thread.java:745)  
unique snapshot  
org.apache.lucene.search.  
FieldValueHitQueue$MultiComparatorsFieldValueHitQueue.lessThan(  
FieldValueHitQueue.java:122)  
org.apache.lucene.search.  
FieldValueHitQueue$MultiComparatorsFieldValueHitQueue.lessThan(  
FieldValueHitQueue.java:99)  
org.apache.lucene.util.PriorityQueue.downHeap(PriorityQueue.java:241)  
org.apache.lucene.util.PriorityQueue.pop(PriorityQueue.java:177)  
org.apache.lucene.search.TopFieldCollector.populateResults(  
TopFieldCollector.java:1203)  
org.apache.lucene.search.TopDocsCollector.topDocs(TopDocsCollector.  
java:156)  
org.apache.lucene.search.TopDocsCollector.topDocs(TopDocsCollector.  
java:93)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:582)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:533)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:510)  
org.apache.lucene.search.IndexSearcher.search(IndexSearcher.java:345)  
org.elasticsearch.search.query.QueryPhase.execute(QueryPhase.java:150  
)  
org.elasticsearch.search.SearchService.executeQueryPhase(  
SearchService.java:275)  
org.elasticsearch.search.action.SearchServiceTransportAction$5.call(  
SearchServiceTransportAction.java:231)  
org.elasticsearch.search.action.SearchServiceTransportAction$5.call(  
SearchServiceTransportAction.java:228)  
org.elasticsearch.search.action.SearchServiceTransportAction$23.run(  
SearchServiceTransportAction.java:559)  
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.  
java:1145)  
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor  
.java:615)  
java.lang.Thread.run(Thread.java:745)

\_nodes/stats?pretty=1  
{  
"cluster\_name" : "elasticsearch",  
"nodes" : {  
"tcr0m37nRAS9appE1pvWLA" : {  
"timestamp" : 1422495166302,  
"name" : "Sepulchre",  
"transport\_address" : "inet[/x.x.x.x:9300]",  
"host" : "sd-70975",  
"ip" : ["inet[/x.x.x.x:9300]", "NONE" ],  
"indices" : {  
"docs" : {  
"count" : 936269,  
"deleted" : 545  
},  
"store" : {  
"size\_in\_bytes" : 856723820,  
"throttle\_time\_in\_millis" : 36  
},  
"indexing" : {  
"index\_total" : 49,  
"index\_time\_in\_millis" : 1058,  
"index\_current" : 0,  
"delete\_total" : 184,  
"delete\_time\_in\_millis" : 84,  
"delete\_current" : 0,  
"noop\_update\_total" : 0,  
"is\_throttled" : false,  
"throttle\_time\_in\_millis" : 0  
},  
"get" : {  
"total" : 0,  
"time\_in\_millis" : 0,  
"exists\_total" : 0,  
"exists\_time\_in\_millis" : 0,  
"missing\_total" : 0,  
"missing\_time\_in\_millis" : 0,  
"current" : 0  
},  
"search" : {  
"open\_contexts" : 16,  
"query\_total" : 11557772,  
"query\_time\_in\_millis" : 91259801,  
"query\_current" : 9,  
"fetch\_total" : 10139269,  
"fetch\_time\_in\_millis" : 37413865,  
"fetch\_current" : 0  
},  
"merges" : {  
"current" : 0,  
"current\_docs" : 0,  
"current\_size\_in\_bytes" : 0,  
"total" : 5,  
"total\_time\_in\_millis" : 5835,  
"total\_docs" : 7258,  
"total\_size\_in\_bytes" : 12852963  
},  
"refresh" : {  
"total" : 287,  
"total\_time\_in\_millis" : 5430  
},  
"flush" : {  
"total" : 45,  
"total\_time\_in\_millis" : 1016  
},  
"warmer" : {  
"current" : 0,  
"total" : 445,  
"total\_time\_in\_millis" : 308  
},  
"filter\_cache" : {  
"memory\_size\_in\_bytes" : 204470568,  
"evictions" : 0  
},  
"id\_cache" : {  
"memory\_size\_in\_bytes" : 0  
},  
"fielddata" : {  
"memory\_size\_in\_bytes" : 6131132,  
"evictions" : 0  
},  
"percolate" : {  
"total" : 0,  
"time\_in\_millis" : 0,  
"current" : 0,  
"memory\_size\_in\_bytes" : -1,  
"memory\_size" : "-1b",  
"queries" : 0  
},  
"completion" : {  
"size\_in\_bytes" : 0  
},  
"segments" : {  
"count" : 679,  
"memory\_in\_bytes" : 41166270,  
"index\_writer\_memory\_in\_bytes" : 0,  
"index\_writer\_max\_memory\_in\_bytes" : 3168274004,  
"version\_map\_memory\_in\_bytes" : 3268,  
"fixed\_bit\_set\_memory\_in\_bytes" : 0  
},  
"translog" : {  
"operations" : 122,  
"size\_in\_bytes" : 17  
},  
"suggest" : {  
"total" : 0,  
"time\_in\_millis" : 0,  
"current" : 0  
},  
"query\_cache" : {  
"memory\_size\_in\_bytes" : 0,  
"evictions" : 0,  
"hit\_count" : 0,  
"miss\_count" : 0  
}  
},  
"os" : {  
"timestamp" : 1422495166542,  
"uptime\_in\_millis" : 10821,  
"load\_average" : [49.66, 47.58, 46.68],  
"cpu" : {  
"sys" : 3,  
"user" : 83,  
"idle" : 12,  
"usage" : 86,  
"stolen" : 0  
},  
"mem" : {  
"free\_in\_bytes" : 221818286080,  
"used\_in\_bytes" : 48629456896,  
"free\_percent" : 84,  
"used\_percent" : 15,  
"actual\_free\_in\_bytes" : 228507299840,  
"actual\_used\_in\_bytes" : 41940443136  
},  
"swap" : {  
"used\_in\_bytes" : 0,  
"free\_in\_bytes" : 10238291968  
}  
},  
"process" : {  
"timestamp" : 1422495166543,  
"open\_file\_descriptors" : 2162,  
"cpu" : {  
"percent" : 982,  
"sys\_in\_millis" : 3025760,  
"user\_in\_millis" : 103162570,  
"total\_in\_millis" : 106188330  
},  
"mem" : {  
"resident\_in\_bytes" : 7449333760,  
"share\_in\_bytes" : 75313152,  
"total\_virtual\_in\_bytes" : 53864665088  
}  
},  
"jvm" : {  
"timestamp" : 1422495166630,  
"uptime\_in\_millis" : 10807489,  
"mem" : {  
"heap\_used\_in\_bytes" : 4758146816,  
"heap\_used\_percent" : 15,  
"heap\_committed\_in\_bytes" : 30894260224,  
"heap\_max\_in\_bytes" : 30894260224,  
"non\_heap\_used\_in\_bytes" : 54495664,  
"non\_heap\_committed\_in\_bytes" : 54984704,  
"pools" : {  
"young" : {  
"used\_in\_bytes" : 536440,  
"max\_in\_bytes" : 1954217984,  
"peak\_used\_in\_bytes" : 1954217984,  
"peak\_max\_in\_bytes" : 1954217984  
},  
"survivor" : {  
"used\_in\_bytes" : 28468776,  
"max\_in\_bytes" : 244252672,  
"peak\_used\_in\_bytes" : 151595664,  
"peak\_max\_in\_bytes" : 244252672  
},  
"old" : {  
"used\_in\_bytes" : 4729141600,  
"max\_in\_bytes" : 28695789568,  
"peak\_used\_in\_bytes" : 4729141600,  
"peak\_max\_in\_bytes" : 28695789568  
}  
}  
},  
"threads" : {  
"count" : 346,  
"peak\_count" : 362  
},  
"gc" : {  
"collectors" : {  
"young" : {  
"collection\_count" : 16969,  
"collection\_time\_in\_millis" : 1384852  
},  
"old" : {  
"collection\_count" : 0,  
"collection\_time\_in\_millis" : 0  
}  
}  
},  
"buffer\_pools" : {  
"direct" : {  
"count" : 1673,  
"used\_in\_bytes" : 70117388,  
"total\_capacity\_in\_bytes" : 70117388  
},  
"mapped" : {  
"count" : 314,  
"used\_in\_bytes" : 154987920,  
"total\_capacity\_in\_bytes" : 154987920  
}  
}  
},  
"thread\_pool" : {  
"generic" : {  
"threads" : 1,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 4,  
"completed" : 1443  
},  
"index" : {  
"threads" : 32,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 32,  
"completed" : 233  
},  
"bench" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"get" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"snapshot" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"merge" : {  
"threads" : 1,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 5,  
"completed" : 116  
},  
"suggest" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"bulk" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"optimize" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"warmer" : {  
"threads" : 1,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 2,  
"completed" : 208  
},  
"flush" : {  
"threads" : 1,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 5,  
"completed" : 113  
},  
"search" : {  
"threads" : 96,  
"queue" : 1,  
"active" : 9,  
"rejected" : 0,  
"largest" : 96,  
"completed" : 24008797  
},  
"listener" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"percolate" : {  
"threads" : 0,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 0,  
"completed" : 0  
},  
"management" : {  
"threads" : 5,  
"queue" : 0,  
"active" : 1,  
"rejected" : 0,  
"largest" : 5,  
"completed" : 63903  
},  
"refresh" : {  
"threads" : 1,  
"queue" : 0,  
"active" : 0,  
"rejected" : 0,  
"largest" : 6,  
"completed" : 112  
}  
},  
"network" : {  
"tcp" : {  
"active\_opens" : 2312428,  
"passive\_opens" : 2484460,  
"curr\_estab" : 582,  
"in\_segs" : 209057765,  
"out\_segs" : 216923023,  
"retrans\_segs" : 49717,  
"estab\_resets" : 1144,  
"attempt\_fails" : 135,  
"in\_errs" : 34,  
"out\_rsts" : 2459  
}  
},  
"fs" : {  
"timestamp" : 1422495166632,  
"total" : {  
"total\_in\_bytes" : 302674501632,  
"free\_in\_bytes" : 264467443712,  
"available\_in\_bytes" : 249068793856,  
"disk\_reads" : 351417,  
"disk\_writes" : 205904,  
"disk\_io\_op" : 557321,  
"disk\_read\_size\_in\_bytes" : 3433067520,  
"disk\_write\_size\_in\_bytes" : 3127025664,  
"disk\_io\_size\_in\_bytes" : 6560093184,  
"disk\_queue" : "0",  
"disk\_service\_time" : "0.1"  
},  
"data" : [ {  
"path" : "/var/lib/elasticsearch/elasticsearch/nodes/0",  
"mount" : "/",  
"dev" : "/dev/sda2",  
"total\_in\_bytes" : 302674501632,  
"free\_in\_bytes" : 264467443712,  
"available\_in\_bytes" : 249068793856,  
"disk\_reads" : 351417,  
"disk\_writes" : 205904,  
"disk\_io\_op" : 557321,  
"disk\_read\_size\_in\_bytes" : 3433067520,  
"disk\_write\_size\_in\_bytes" : 3127025664,  
"disk\_io\_size\_in\_bytes" : 6560093184,  
"disk\_queue" : "0",  
"disk\_service\_time" : "0.1"  
} ]  
},  
"transport" : {  
"server\_open" : 13,  
"rx\_count" : 6,  
"rx\_size\_in\_bytes" : 1380,  
"tx\_count" : 6,  
"tx\_size\_in\_bytes" : 1380  
},  
"http" : {  
"current\_open" : 11,  
"total\_opened" : 2311818  
},  
"breakers" : {  
"request" : {  
"limit\_size\_in\_bytes" : 12357704089,  
"limit\_size" : "11.5gb",  
"estimated\_size\_in\_bytes" : 16440,  
"estimated\_size" : "16kb",  
"overhead" : 1.0,  
"tripped" : 0  
},  
"fielddata" : {  
"limit\_size\_in\_bytes" : 18536556134,  
"limit\_size" : "17.2gb",  
"estimated\_size\_in\_bytes" : 6131132,  
"estimated\_size" : "5.8mb",  
"overhead" : 1.03,  
"tripped" : 0  
},  
"parent" : {  
"limit\_size\_in\_bytes" : 21625982156,  
"limit\_size" : "20.1gb",  
"estimated\_size\_in\_bytes" : 6147572,  
"estimated\_size" : "5.8mb",  
"overhead" : 1.0,  
"tripped" : 0  
}  
}  
}  
}  
}

On Wednesday, January 28, 2015 at 11:41:16 PM UTC+1, Oto Iashvili wrote:

> Hi,
> 
> thanks a lot for answer,  
> Ive tried several value for heap, between 26 and 32, but I didnt see any  
> difference.
> 
> I remove G1 and put back default parameter. But still same pb.  
> I was saying around 500%, but it is just average, i goes sometimes up to  
> 2000%
> 
> I was also thinking to take several server, but right now it is not  
> possible.
> 
> I was using a smaller server just before, with 96go of ram and I was  
> working better. I trie to put same parameters as before, but it is not much  
> slower.
> 
> On Wednesday, January 28, 2015 at 2:36:29 PM UTC+1, Jilles van Gurp wrote:
> 
> > How much heap are you giving to ES? With this many requests, if your  
> > setup is not falling over it is probably not garbage collect related  
> > because that would result in very noticable delays/unavailability of es.  
> > 32GB should be a good value given how much memory you have. Also, you  
> > probably want to use doc\_values in your mapping so that you can utilize the  
> > os file cache and move some of the memory pressure on the heap. You seem to  
> > have plenty of ram, so your entire dataset should easily fit in RAM.
> > 
> > Also, don't use G1 for elasticsearch. There are known issues with that  
> > particular garbage collector in combination with lucene. CMS is the best  
> > option for ES.
> > 
> > 500% of 20 cores doesn't sound that bad; you'd max them out at 4000%.  
> > Still, it would be nice to know what it is doing. In any case, you might  
> > want to try out marvel to find out where your setup is bottlenecked. Also,  
> > you might want to consider scaling horizontally instead of vertically. Many  
> > smaller servers can be nicer than one big one.
> > 
> > On Wednesday, January 28, 2015 at 2:09:22 PM UTC+1, Oto Iashvili wrote:
> > 
> > > Hi
> > > 
> > > I have a website for classified. For this I'm using elasticsearch,  
> > > postgres and rails on a same ubuntu 14.04 dedicated server, with 256go of  
> > > RAM and 20 cores, 40 threads .
> > > 
> > > I have 10 indexes on elasticsearch, each have default numbers of shardes  
> > > (5). They have between 1000 and 400 000 classifieds dependings on which  
> > > index.  
> > > approximatly 5000 requests per minute, 2/3 making an elasticsearch  
> > > request.
> > > 
> > > according to htop, jvm is using around 500% of CPU  
> > > I try different options, I reduce number of shardes per index, I also  
> > > try to change JAVA\_OPTS of followed
> > > 
> > > ```
> > > #JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
> > > #JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
> > > 
> > > #JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
> > > #JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
> > > 
> > > JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
> > > 
> > > ```
> > > 
> > > but it doesnt seems to change anything.
> > > 
> > > so to questions :
> > > 
> > > - when you change any setting on elasticsearch, and then restart,  
> > > should the improvement (if any) be visible immediatly or can it arrive a  
> > > bit later thanks to cache or any thing else ?
> > > - can any one help me to find good configuration for JVM / elasticsearch  
> > > so it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/d1b4c785-d6d0-4677-a319-1dc9ea00c6da%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/d1b4c785-d6d0-4677-a319-1dc9ea00c6da%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![Arie](https://avatars.discourse-cdn.com/v4/letter/a/c2a13f/32.png) [@Arie](https://discuss.elastic.co/u/Arie)\
**Post date:** [January 29, 2015, 9:02am UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/5 "2015-01-29T09:02:00Z")

</div>

Just an idea.

You could try running two ES instances as a cluster on one machine if there  
is no other option.

On Wednesday, January 28, 2015 at 2:09:22 PM UTC+1, Oto Iashvili wrote:

> Hi
> 
> I have a website for classified. For this I'm using elasticsearch,  
> postgres and rails on a same ubuntu 14.04 dedicated server, with 256go of  
> RAM and 20 cores, 40 threads .
> 
> I have 10 indexes on elasticsearch, each have default numbers of shardes  
> (5). They have between 1000 and 400 000 classifieds dependings on which  
> index.  
> approximatly 5000 requests per minute, 2/3 making an elasticsearch request.
> 
> according to htop, jvm is using around 500% of CPU  
> I try different options, I reduce number of shardes per index, I also try  
> to change JAVA\_OPTS of followed
> 
> ```
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
> 
> #JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
> #JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
> 
> JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
> 
> ```
> 
> but it doesnt seems to change anything.
> 
> so to questions :
> 
> - when you change any setting on elasticsearch, and then restart, should  
> the improvement (if any) be visible immediatly or can it arrive a bit later  
> thanks to cache or any thing else ?
> - can any one help me to find good configuration for JVM / elasticsearch  
> so it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/ef4bad2e-42c7-4450-ba95-2da27f5a6b58%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/ef4bad2e-42c7-4450-ba95-2da27f5a6b58%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![oto\_iashvili](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@oto\_iashvili](https://discuss.elastic.co/u/oto_iashvili)\
**Post date:** [January 29, 2015, 10:38am UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/6 "2015-01-29T10:38:22Z")

</div>

why not ? Could u tell me how to do such ? and also explain why will it be  
better ?

thanks a lot for your help

On Thursday, January 29, 2015 at 10:02:00 AM UTC+1, Arie wrote:

> Just an idea.
> 
> You could try running two ES instances as a cluster on one machine if  
> there is no other option.
> 
> On Wednesday, January 28, 2015 at 2:09:22 PM UTC+1, Oto Iashvili wrote:
> 
> > Hi
> > 
> > I have a website for classified. For this I'm using elasticsearch,  
> > postgres and rails on a same ubuntu 14.04 dedicated server, with 256go of  
> > RAM and 20 cores, 40 threads .
> > 
> > I have 10 indexes on elasticsearch, each have default numbers of shardes  
> > (5). They have between 1000 and 400 000 classifieds dependings on which  
> > index.  
> > approximatly 5000 requests per minute, 2/3 making an elasticsearch  
> > request.
> > 
> > according to htop, jvm is using around 500% of CPU  
> > I try different options, I reduce number of shardes per index, I also try  
> > to change JAVA\_OPTS of followed
> > 
> > ```
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseParNewGC"
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseConcMarkSweepGC"
> > 
> > #JAVA_OPTS="$JAVA_OPTS -XX:CMSInitiatingOccupancyFraction=75"
> > #JAVA_OPTS="$JAVA_OPTS -XX:+UseCMSInitiatingOccupancyOnly"
> > 
> > JAVA_OPTS="$JAVA_OPTS -XX:+UseG1GC"
> > 
> > ```
> > 
> > but it doesnt seems to change anything.
> > 
> > so to questions :
> > 
> > - when you change any setting on elasticsearch, and then restart, should  
> > the improvement (if any) be visible immediatly or can it arrive a bit later  
> > thanks to cache or any thing else ?
> > - can any one help me to find good configuration for JVM / elasticsearch  
> > so it will not take that many ressources

--  
You received this message because you are subscribed to the Google Groups "elasticsearch" group.  
To unsubscribe from this group and stop receiving emails from it, send an email to [elasticsearch+unsubscribe@googlegroups.com](mailto:elasticsearch+unsubscribe@googlegroups.com).  
To view this discussion on the web visit [https://groups.google.com/d/msgid/elasticsearch/9db68f64-e79d-4592-9085-0633eec7360f%40googlegroups.com](https://groups.google.com/d/msgid/elasticsearch/9db68f64-e79d-4592-9085-0633eec7360f%40googlegroups.com).  
For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout).

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [July 6, 2017, 12:36am UTC](https://discuss.elastic.co/t/optimize-elasticsearch-jvm/21874/7 "2017-07-06T00:36:00Z")

</div>