@Mark, I'm using this cluster only for logstash, so I have 1 new index 
p/day, each index has between 30k and 100k documents. I use curator in a 
crontab to close old logs older than X days so I don't have them all loaded 
into memory at the same time. Right now I have 8 indices opened with a 
total of ~500k documents occupying ~4.5Gb of disk space. I read a bit about 
OOM with ElasticSearch and I couldn't figure out my problem yet. I followed 
ES docs memory tips:

http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/setup-configuration.html

And I see that mlockall is set to true on both instances.

@Ivan, yes, ES and Logstash are the only processes running on the servers. 
No, I didn't set direct memory values. Should I? If so, which sort of 
values and how do you determine those?

It's been a while since I worked with JVMs so my knowledge on Java Memory 
stuff is kind of rusty :). Thanks for the help!

On Wednesday, July 16, 2014 5:40:27 AM UTC+2, Ivan Brusic wrote:
>
> Direct memory is off heap memory. Are elasticsearch and logstash the only 
> processes on those servers? Did you set an explicit direct memory value?
>
> -- 
> Ivan
> On Jul 15, 2014 3:46 PM, "Mark Walkom" <[email protected] 
> <javascript:>> wrote:
>
>> How much data do you have in ES, index count and total size of all 
>> indexes?
>>
>> OutOfMemoryError means you ran out of heap, which could mean a few things.
>>
>> Regards,
>> Mark Walkom
>>
>> Infrastructure Engineer
>> Campaign Monitor
>> email: [email protected] <javascript:>
>> web: www.campaignmonitor.com
>>
>>
>> On 15 July 2014 22:33, Pedro Jerónimo <[email protected] <javascript:>
>> > wrote:
>>
>>> I have a cluster of 2 ES machines with a lot of indexing, not so much 
>>> searching. I'm using 2 EC2 machines with 30gb of RAM and I'm running ES on 
>>> each with 12gb heap (ES_HEAP_SIZE) and one of them (let's call it 
>>> logs1) is running logstash as well, with 2gb heap. The master node is logs1 
>>> and the other instance is logs2. I start the cluster and every things looks 
>>> fine, but after a while (1-3 days) I get the following error on logs1:
>>>
>>> [2014-07-15 12:26:39,867][WARN ][transport.netty ] [Keen Marlow] 
>>> exception caught on transport layer [[id: 0x16801a48, /XX.XX.XXX.XX:36314 
>>> => /XX.XXX.XX.XX:9300]], closing connection 
>>> java.lang.OutOfMemoryError: Direct buffer memory
>>> ...Stack Trace...
>>>
>>> And then the cluster is no longer connected and if I try to restart 
>>> logs2, I get the same error above for logs1 and this one for logs2:
>>>
>>> [2014-07-15 12:27:39,282][INFO ][discovery.ec2 ] [Betty Ross Banner] 
>>> failed to send join request to master [[Keen 
>>> Marlow][9a7FIRpBSrKQcdcV_sjSTw][ip-XX-XX-XXX-XX][inet[/XX.XX.XXX.XX:9300]]{aws_availability_zone=us-west-2a,
>>>  
>>> master=true}], reason 
>>> [org.elasticsearch.transport.RemoteTransportException: [Keen 
>>> Marlow][inet[/XX.XX.XXX.XX:9300]][discovery/zen/join]; 
>>> org.elasticsearch.transport.NodeDisconnectedException: [Betty Ross 
>>> Banner][inet[/XX.XXX.XX.XX:9300]][discovery/zen/join/validate] disconnected]
>>>
>>> Is there any memory configuration I should tune up a bit? I'm kind of 
>>> new to ElasticSearch so I'd love some help! :).
>>>
>>> Thanks!
>>>
>>> Pedro
>>>  
>>> -- 
>>> 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 [email protected] <javascript:>.
>>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/elasticsearch/7d155e16-c8ff-40b0-ac0a-e3719e8296c3%40googlegroups.com
>>>  
>>> <https://groups.google.com/d/msgid/elasticsearch/7d155e16-c8ff-40b0-ac0a-e3719e8296c3%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
>>  -- 
>> 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 [email protected] <javascript:>.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/elasticsearch/CAEM624Yb9mhh3-E9jYvFFaY_dB8rF-R5yZCGYCHrvWNsur-MwQ%40mail.gmail.com
>>  
>> <https://groups.google.com/d/msgid/elasticsearch/CAEM624Yb9mhh3-E9jYvFFaY_dB8rF-R5yZCGYCHrvWNsur-MwQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>

-- 
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 [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/bb843355-0c29-4ec1-803a-9149b89fc36a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to