It looks like heap pressure. How many indices, how many shards, how much data do you have in the cluster?
On 8 March 2015 at 19:24, Chris Neal <[email protected]> wrote: > Thank you Mark for your reply. > > I do have Marvel running, on a separate cluster even, so I do have that > data from the time of the problem. I've attached 4 screenshots for > reference. > > It appears that node 10.0.0.12 (the green line on the charts) had issues. > The heap usage drops from 80% to 0%. I'm guessing that is some sort of > crash, because the heap should not empty itself. Also its load goes to 0. > > I also see a lot of Old GC duration on 10.0.0.45 (blue line). Lots of > excessive Old GC Counts, so it does appear that the problem was memory > pressure on this node. That's what I was thinking, but was hoping for > validation on that. > > If it was, I'm hoping to get some suggestions on what to do about it. As > I mentioned in the original post, I've tweaked I think needs tweaking based > on the system, and it still happens. > > Maybe it's just that I'm pushing the cluster too much for the resources > I'm giving it, and it "just won't work". > > The index rate was only about 2500/sec, and the search request rate had > one small spike that went to 3.0. But 3 searches in one timeslice is > nothing. > > Thanks again for the help and reading all this stuff. It is appreciated. > Hopefully I can get a solution to keep the cluster stable. > > Chris > > On Fri, Mar 6, 2015 at 3:01 PM, Mark Walkom <[email protected]> wrote: > >> You really need some kind of monitoring, like Marvel, around this to give >> you an idea of what was happening prior to the OOM. >> Generally a node becoming unresponsive will be due to GC, so take a look >> at the timings there. >> >> On 5 March 2015 at 02:32, Chris Neal <[email protected]> wrote: >> >>> Hi all, >>> >>> I'm hoping someone can help me piece together the below log >>> entries/stack traces/Exceptions. I have a 3 node cluster in Development in >>> EC2, and two of them had issues. I'm running ES 1.4.4, 32GB RAM, 16GB >>> heaps, dedicated servers to ES. My idex rate averages about 10k/sec. >>> There were no searches going on at the time of the incident. >>> >>> It appears to me that node 10.0.0.12 began timing out requests to >>> 10.0.45, indicating that 10.0.0.45 was having issues. >>> Then at 4:36, 10.0.0.12 logs the ERROR about "Uncaught exception: >>> IndexWriter already closed", caused by an OOME. >>> Then at 4:43, 10.0.0.45 hits the "Create failed" WARN, and logs an OOME. >>> >>> Then things are basically down and unresponsive. >>> >>> What is weird to me is that if 10.0.0.45 was the node having issues, why >>> did 10.0.0.12 log an exception 7 minutes before that? Did both nodes run >>> out of memory? Or is one of the Exceptions actually saying, "I see that >>> this other node hit an OOME, and I'm telling you about it." >>> >>> I have a few values tweaked in the elasticsearch.yml file to try and >>> keep this from happening (configured from Puppet): >>> >>> 'indices.breaker.fielddata.limit' => '20%', >>> 'indices.breaker.total.limit' => '25%', >>> 'indices.breaker.request.limit' => '10%', >>> 'index.merge.scheduler.type' => 'concurrent', >>> 'index.merge.scheduler.max_thread_count' => '1', >>> 'index.merge.policy.type' => 'tiered', >>> 'index.merge.policy.max_merged_segment' => '1gb', >>> 'index.merge.policy.segments_per_tier' => '4', >>> 'index.merge.policy.max_merge_at_once' => '4', >>> 'index.merge.policy.max_merge_at_once_explicit' => '4', >>> 'indices.memory.index_buffer_size' => '10%', >>> 'indices.store.throttle.type' => 'none', >>> 'index.translog.flush_threshold_size' => '1GB', >>> >>> I have done a fair bit of reading on this, and have tried about >>> everything I can think of. :( >>> >>> Can anyone tell me what caused this scenario, and what can be done to >>> avoid it? >>> Thank you so much for taking the time to read this. >>> Chris >>> >>> ===== >>> *On server 10.0.0.12 <http://10.0.0.12>:* >>> >>> [2015-03-04 03:56:12,548][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [20456ms] ago, timed out [5392ms] ago, action >>> [cluster:monitor/nodes/st >>> ats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70061596] >>> [2015-03-04 04:06:02,407][INFO ][index.engine.internal ] >>> [elasticsearch-ip-10-0-0-12] [derbysoft-ihg-20150304][2] now throttling >>> indexing: numMergesInFlight=4, maxNumMerges=3 >>> [2015-03-04 04:06:04,141][INFO ][index.engine.internal ] >>> [elasticsearch-ip-10-0-0-12] [derbysoft-ihg-20150304][2] stop throttling >>> indexing: numMergesInFlight=2, maxNumMerges=3 >>> [2015-03-04 04:12:26,194][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [15709ms] ago, timed out [708ms] ago, action >>> [cluster:monitor/nodes/sta >>> ts[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70098828] >>> [2015-03-04 04:23:40,778][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [21030ms] ago, timed out [6030ms] ago, action >>> [cluster:monitor/nodes/st >>> ats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70124234] >>> [2015-03-04 04:24:47,023][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [27275ms] ago, timed out [12275ms] ago, action >>> [cluster:monitor/nodes/s >>> tats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70126273] >>> [2015-03-04 04:25:39,180][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [19431ms] ago, timed out [4431ms] ago, action >>> [cluster:monitor/nodes/st >>> ats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70127835] >>> [2015-03-04 04:26:40,775][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [19241ms] ago, timed out [4241ms] ago, action >>> [cluster:monitor/nodes/st >>> ats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70129981] >>> [2015-03-04 04:27:14,329][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [22676ms] ago, timed out [6688ms] ago, action >>> [cluster:monitor/nodes/stats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70130668] >>> [2015-03-04 04:28:15,695][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [24042ms] ago, timed out [9041ms] ago, action >>> [cluster:monitor/nodes/stats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70132644] >>> [2015-03-04 04:29:38,102][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [16448ms] ago, timed out [1448ms] ago, action >>> [cluster:monitor/nodes/stats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70135333] >>> [2015-03-04 04:33:42,393][WARN ][transport ] >>> [elasticsearch-ip-10-0-0-12] Received response for a request that has timed >>> out, sent [20738ms] ago, timed out [5737ms] ago, action >>> [cluster:monitor/nodes/stats[n]], node >>> [[elasticsearch-ip-10-0-0-45][i4gmsxs0Q0eyvPWjajNV5A][ip-10-0-0-45.us-west-2.compute.internal][inet[ip-10-0-0-45.us-west-2.compute.internal/10.0.0.45:9300]]{master=true}], >>> id [70142427] >>> [2015-03-04 04:36:08,788][ERROR][marvel.agent ] >>> [elasticsearch-ip-10-0-0-12] Background thread had an uncaught exception: >>> org.apache.lucene.store.AlreadyClosedException: this IndexWriter is >>> closed >>> at >>> org.apache.lucene.index.IndexWriter.ensureOpen(IndexWriter.java:698) >>> at >>> org.apache.lucene.index.IndexWriter.ensureOpen(IndexWriter.java:712) >>> at >>> org.apache.lucene.index.IndexWriter.ramBytesUsed(IndexWriter.java:462) >>> at >>> org.elasticsearch.index.engine.internal.InternalEngine.segmentsStats(InternalEngine.java:1224) >>> at >>> org.elasticsearch.index.shard.service.InternalIndexShard.segmentStats(InternalIndexShard.java:555) >>> at >>> org.elasticsearch.action.admin.indices.stats.CommonStats.<init>(CommonStats.java:170) >>> at >>> org.elasticsearch.action.admin.indices.stats.ShardStats.<init>(ShardStats.java:49) >>> at >>> org.elasticsearch.indices.InternalIndicesService.stats(InternalIndicesService.java:212) >>> at >>> org.elasticsearch.indices.InternalIndicesService.stats(InternalIndicesService.java:172) >>> at >>> org.elasticsearch.node.service.NodeService.stats(NodeService.java:138) >>> at >>> org.elasticsearch.marvel.agent.AgentService$ExportingWorker.exportNodeStats(AgentService.java:300) >>> at >>> org.elasticsearch.marvel.agent.AgentService$ExportingWorker.run(AgentService.java:225) >>> at java.lang.Thread.run(Thread.java:745) >>> Caused by: java.lang.OutOfMemoryError: Java heap space >>> >>> ===== >>> *On server 10.0.0.45 <http://10.0.0.45>:* >>> >>> [2015-03-04 04:43:27,245][WARN ][index.engine.internal ] >>> [elasticsearch-ip-10-0-0-45] [myindex-20150304][1] failed engine >>> [indices:data/write/bulk[s] failed on replica] >>> org.elasticsearch.index.engine.CreateFailedEngineException: >>> [myindex-20150304][1] Create failed for [my_type#AUvjGHoiku-fZf277h_4] >>> at >>> org.elasticsearch.index.engine.internal.InternalEngine.create(InternalEngine.java:421) >>> at >>> org.elasticsearch.index.shard.service.InternalIndexShard.create(InternalIndexShard.java:403) >>> at >>> org.elasticsearch.action.bulk.TransportShardBulkAction.shardOperationOnReplica(TransportShardBulkAction.java:595) >>> at >>> org.elasticsearch.action.support.replication.TransportShardReplicationOperationAction$ReplicaOperationTransportHandler.messageReceived(TransportShardReplicationOperationAction.java:246) >>> at >>> org.elasticsearch.action.support.replication.TransportShardReplicationOperationAction$ReplicaOperationTransportHandler.messageReceived(TransportShardReplicationOperationAction.java:225) >>> at >>> org.elasticsearch.transport.netty.MessageChannelHandler$RequestHandler.run(MessageChannelHandler.java:275) >>> at >>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) >>> at >>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) >>> at java.lang.Thread.run(Thread.java:745) >>> Caused by: org.apache.lucene.store.AlreadyClosedException: this >>> IndexWriter is closed >>> at >>> org.apache.lucene.index.IndexWriter.ensureOpen(IndexWriter.java:698) >>> at >>> org.apache.lucene.index.IndexWriter.ensureOpen(IndexWriter.java:712) >>> at >>> org.apache.lucene.index.IndexWriter.updateDocument(IndexWriter.java:1507) >>> at >>> org.apache.lucene.index.IndexWriter.addDocument(IndexWriter.java:1246) >>> at >>> org.elasticsearch.index.engine.internal.InternalEngine.innerCreateNoLock(InternalEngine.java:502) >>> at >>> org.elasticsearch.index.engine.internal.InternalEngine.innerCreate(InternalEngine.java:444) >>> at >>> org.elasticsearch.index.engine.internal.InternalEngine.create(InternalEngine.java:413) >>> ... 8 more >>> Caused by: java.lang.OutOfMemoryError: Java heap space >>> >>> ===== >>> >>> >>> >>> -- >>> 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/CAND3DphzaT3Np5TBW%2B-h_aOo9BScPu_5QO9qCqnYLp__JCjOPA%40mail.gmail.com >>> <https://groups.google.com/d/msgid/elasticsearch/CAND3DphzaT3Np5TBW%2B-h_aOo9BScPu_5QO9qCqnYLp__JCjOPA%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/CAEYi1X_9uMAwF7nkZRnDvB9DAMmkSGrNG1HiWWvNgTRcg2TM8w%40mail.gmail.com >> <https://groups.google.com/d/msgid/elasticsearch/CAEYi1X_9uMAwF7nkZRnDvB9DAMmkSGrNG1HiWWvNgTRcg2TM8w%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/CAND3Dpjr5ZUHKeWZROCJ6uCCjmEU3_geDuSdK96_-uqL6qGX2A%40mail.gmail.com > <https://groups.google.com/d/msgid/elasticsearch/CAND3Dpjr5ZUHKeWZROCJ6uCCjmEU3_geDuSdK96_-uqL6qGX2A%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/CAEYi1X88-hv1vp3xwJsz2kPex3tAND-rx%3DT-CEO1GXO0CkwSww%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
