Question

在logstash 2.4上使用Ubuntu 16.04（我有理由）

root@logbox:/etc/logstash/conf.d# ls -al
total 16
drwxrwxr-x 2 root root 4096 Nov  2 19:53 .
drwxrwxr-x 3 root root 4096 Nov  2 15:46 ..
-rwxrwxrwx 1 root root  277 Nov  2 19:52 01_01_input.conf
-rwxrwxrwx 1 root root  604 Nov  2 19:48 03_02_output_pa_http.conf


root@logbox:/etc/logstash/conf.d# cat *.conf

input {

  stdin {}

  file {
    path => "/usr/share/logstash/files/production_input.txt"
    start_position => "beginning"
    # codec => plain { charset => "ISO-8859-1" }
    codec => json
    # add_field => [ 'redis_db', '10' ]
  }



}
output {

  stdout {}

}


root@logbox:/etc/logstash/conf.d# ls -al /usr/share/logstash/files/production_input.txt
-rwxrwxrwx 1 root root 11910 Nov  2 16:09 /usr/share/logstash/files/production_input.txt

但是...

vagrant@logbox:/etc/logstash/conf.d$ sudo tail -f /var/log/logstash/logstash*


Sending logstash logs to /var/log/logstash/logstash.log.

==> /var/log/logstash/logstash.log <==
{:timestamp=>"2018-11-02T19:59:52.947000+0000", :message=>"Pipeline main started"}

==> /var/log/logstash/logstash.stdout <==
{:timestamp=>"2018-11-02T19:59:52.947000+0000", :message=>"Pipeline main started"}

文件未打印在stdout中...我已经尝试了两种编解码器（json和plain）

编辑：实际上可能是文件一次只读的情况；这和sincedb有关系吗？如何强制Logstash重新读取？ start_position => beginning还不够吗？

Answer 1

start_position在第一次读取文件时使用。

如果要强制logstash再次读取它，则需要在文件输入插件conf中添加以下参数： sincedb_path => "/dev/null"

请阅读以获取更多说明 How to force Logstash to reparse a file?

Logstash：最简单的管道可能无法正常工作

1 个答案: