Problems to run elasticsearch


(Jan Tappenbeck) #1

Hi !

we are a little company in german and tested to use a local wiki for documentations. i have install on win7 64bit machine local a mediawiki and elasticsearch. i try to install one year ago - without the nice help in web(inar).

out main-it now change to move the machine from a physical to a virtual machine.

the mediawiki will run without problems on the new machine. but elasticsearch had problems!

when the service runs in combination of cirrussearch for mediawiki the logfile gets many reports so the file will will be more and more (GB+) and the pc-performance goes to sleep!

Here is a part of the Logfile at beginning:

[2015-12-16 15:06:10,251][INFO ][node ] [Mangle] version[1.4.4], pid[3972], build[c88f77f/2015-02-19T13:05:36Z]
[2015-12-16 15:06:10,252][INFO ][node ] [Mangle] initializing ...
[2015-12-16 15:06:10,266][INFO ][plugins ] [Mangle] loaded [], sites []
[2015-12-16 15:06:17,145][INFO ][node ] [Mangle] initialized
[2015-12-16 15:06:17,147][INFO ][node ] [Mangle] starting ...
[2015-12-16 15:06:17,480][INFO ][transport ] [Mangle] bound_address {inet[/0:0:0:0:0:0:0:0:9300]}, publish_address {inet[/10.32.7.124:9300]}
[2015-12-16 15:06:17,615][INFO ][discovery ] [Mangle] elasticsearch/aBHurpT1SaGzBFybtqkleA
[2015-12-16 15:06:21,411][INFO ][cluster.service ] [Mangle] new_master [Mangle][aBHurpT1SaGzBFybtqkleA][MA22-Wiki-001][inet[/10.32.7.124:9300]], reason: zen-disco-join (elected_as_master)
[2015-12-16 15:06:21,536][INFO ][http ] [Mangle] bound_address {inet[/0:0:0:0:0:0:0:0:9200]}, publish_address {inet[/10.32.7.124:9200]}
[2015-12-16 15:06:21,537][INFO ][node ] [Mangle] started
[2015-12-16 15:06:23,688][INFO ][gateway ] [Mangle] recovered [4] indices into cluster_state
[2015-12-16 15:06:24,095][WARN ][indices.cluster ] [Mangle] [eblwiki-eblw__general_first][0] failed to start shard
org.elasticsearch.index.gateway.IndexShardGatewayRecoveryException: [eblwiki-eblw__general_first][0] failed to fetch index version after copying it over
at org.elasticsearch.index.gateway.local.LocalIndexShardGateway.recover(LocalIndexShardGateway.java:158)
at org.elasticsearch.index.gateway.IndexShardGatewayService$1.run(IndexShardGatewayService.java:132)
at java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
at java.lang.Thread.run(Unknown Source)
Caused by: org.apache.lucene.index.CorruptIndexException: [eblwiki-eblw__general_first][0] Preexisting corrupted index [corrupted_HXhLgNexRJevJjm02WLiDw] caused by: CorruptIndexException[codec footer mismatch: actual footer=0 vs expected footer=-1071082520 (resource: MMapIndexInput(path="C:\elasticsearch\data\elasticsearch\nodes\0\indices\eblwiki-eblw__general_first\0\index_je.tvd"))]
org.apache.lucene.index.CorruptIndexException: codec footer mismatch: actual footer=0 vs expected footer=
...

could some one help with a good tipp to make elasticsearch fit for use....

my english is not very well, so it would be nice to get answer by german workers.

current the system must work like it is - in feature we will make the system better and it could be possible to get prof. help.

regards Jan


(Alexander Reelsen) #2

Hey,

did you upgrade from an older Elasticsearch version? is it possible that this data was already corrupted and cannot be loaded anymore? Elasticsearch 1.4 and above added a lot of checksumming compared to prior versions which potential can result in data being shown as corrupted, that has gone undetected before.

Another cause might be that you ran out of disk space so that Elasticsearch could not write data, but thats just a far fetched assumption.

Can you reindex the whole data?
Can you reproduce this problem continously, even with a clean empty index?
Are there OutOfMemory Exceptions in your logs?
Can you explain in more detail what you mean with the performance issue? what happens in order for this to occur?

--Alex


(system) #3