Error inserting data

Elastic Stack Elasticsearch
1

Since the upgrade to 6.0.0, we see this error:

java.lang.IllegalArgumentException: Rejecting mapping update to [logstash-2017.11.22] as the final mapping would have more than 1 type: [logfile, syslog]

Is "type" een reserved field which can have only 1 value in an index?

The record that fails:

{"path":"/xxx/http_access.log","@timestamp":"2017-11-22T12:03:40.149Z",
"log_indexer":"yyy","log_source":"washttpaccess","@version":"1","host":"xxx","log_shipper":"xxx",
"message":"10.aa.bb.41 - [22/Nov/2017:13:03:39 +0100] HEAD /dddd HTTP/1.1 401 - 0 - -\r","type":"logfile","tags":["kafka"]}]}]

2

Probably look at the template. Probably there is a template for both types which is not supported anymore.

1 Like
3

I'm seeing something issues with our update. It looks like overnight is when our events stopped processing (matching up with a new index) We see the following elasticsearch error:

[ciscoswitch-2017.11.22][3] failed to execute bulk item (index) BulkShardRequest [[ciscoswitch-2017.11.22][3]] containing [index {[ciscoswitch-2017.11.22][doc][94r45F8BPUaikWKCqLiK], source[{"cisco_type":"LINK-3-UPDOWN","message_type_id":"187","description":"Interface","message":"<187>707219: Nov 22 12:23:42.202 CST: %LINK-3-UPDOWN: Interface GigabitEthernet1/0/15, changed state to down","interface":"GigabitEthernet1/0/15","tags":["cisco-switch","cisco-switch-processed","30-filter-ciscoSwitch"],"cisco_time":"Nov 22 12:23:42.202","@timestamp":"2017-11-22T18:23:42.210Z","event_no":"707219","@version":"1","host":"192.168.52.4","eventTime":"2017-11-22T18:23:42.202Z","status":"changed state to down"}]}]
java.lang.IllegalArgumentException: Rejecting mapping update to [ciscoswitch-2017.11.22] as the final mapping would have more than 1 type: [cisco-switch, doc]

But nowhere are we defining the doc type within our config. Also the type "cisco-switch" has been shifted to tag but it still looks like it is lingering from somewhere.

Is there a way to prevent the new doc type from being automatically added to events in v6?

4

I think that doc is the new default value when you don't explicitly define a type.
Is your data coming from logstash?

What gives GET /_template on your system?

5

elastic support came with this addition to the elasticsearch-output statement in the logstash config:

document_type => doc

The setting will get active for new created indices.

It solved my issue in the test environment, production is waiting for an index rollover.

6

I encountered the same problem, I tried your solution, but it seems no effect, can be described in more detail please?，thanks。

7

The issue pertaining to a few of our indices resolved itself when the index rolled over the following day, but we are still seeing this with data originating via filebeat.

Our data originates on a remote machine and is shipped using filebeat into logstash with a beats input. The filebeat prospector is pretty basic and follows the documentation online, aside from the path:

filebeat.prospectors:
- type: log
  paths:
    - /Library/Log/login.log

In our logstash dead letter queue we see the following:

[����MJ72017-11-27T15:54:06.683Z�dMETA�dbeathfilebeatjip_addressn192.168.43.207dtypeclog�dDATA�foffsetZ�bipn192.168.43.207jinput_typeclogfsourcex/Library/Logs/login.loggmessagex=LOGIN,USERNAME,Mon Nov 27 09:53:35 CST 2017,192.168.43.207dtypeclogdtags�hmacLoginxbeats_input_codec_plain_appliedx19-filter-computerLoginsj@timestampx2017-11-27T15:53:35.000Zkreceived_atx2017-11-27T15:53:43.095Zh@versiona1dbeat�dnamelC2017025-L04hhostnameilocalhostgversione5.6.3�dhostilocalhostieventTimexMon Nov 27 09:53:35 CST 2017factioneLOGINelasticsearch@388aa477a7790c221ae6013b8ba2d79913e2a3b659e75668880e12ccc8f7f1ee�Could not index event to Elasticsearch. status: 400, action: ["index", {:_id=>nil, :_index=>"login-2017.11.27", :_type=>"log", :_routing=>nil}, #<LogStash::Event:0x2fbf0773>], response: {"index"=>{"_index"=>"login-2017.11.27", "_type"=>"log", "_id"=>"JLUv_l8BJvIoQpoCY3Zl", "status"=>400, "error"=>{"type"=>"illegal_argument_exception", "reason"=>"Rejecting mapping update to [login-2017.11.27] as the final mapping would have more than 1 type: [log, type1]"}}}

If I look at our index template:

{
  "login" : {
    "order" : 0,
    "index_patterns" : [
      "login-*"
    ],
    "settings" : {
      "index" : {
        "number_of_shards" : "1"
      }
    },
    "mappings" : {
      "type1" : {
        "_source" : {
          "enabled" : false
        },
        "properties" : {
          "host_name" : {
            "type" : "keyword"
          },
          "eventTime" : {
            "type" : "date",
            "format" : "EEE MMM dd HH:mm:ss z YYYY"
          }
        }
      }
    },
    "aliases" : { }
  }
}

Would I just need to replace type1 in the index template mappings with log for the event to properly store in the index? I am working toward removing types and working on tags within our configuration to be more future proof.

8

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.