fluent/fluentd · error · Fluent::ConfigError

Invalid <secondary> section for non-buffered plugin

Error message

Invalid <secondary> section for non-buffered plugin

What it means

A `<secondary>` section was configured under an output plugin that does not buffer at all (`@buffering` is false, i.e. the plugin only implements synchronous `#process`). Secondary output exists solely to receive buffered chunks after retry limits, so it is meaningless — and rejected — for non-buffered outputs, which have no retry queue.

Source

Thrown at lib/fluent/plugin/output.rb:401

              end
            end
          end

          if (@flush_mode != :interval) && buffer_conf.has_key?('flush_interval')
            if buffer_conf.has_key?('flush_mode')
              raise Fluent::ConfigError, "'flush_interval' can't be specified when 'flush_mode' is not 'interval' explicitly: '#{@flush_mode}'"
            else
              log.warn "'flush_interval' is ignored because default 'flush_mode' is not 'interval': '#{@flush_mode}'"
            end
          end

          if @buffer.queued_chunks_limit_size.nil?
            @buffer.queued_chunks_limit_size = @buffer_config.flush_thread_count
          end
        end

        if @secondary_config
          raise Fluent::ConfigError, "Invalid <secondary> section for non-buffered plugin" unless @buffering
          raise Fluent::ConfigError, "<secondary> section cannot have <buffer> section" if @secondary_config.buffer
          raise Fluent::ConfigError, "<secondary> section cannot have <secondary> section" if @secondary_config.secondary
          if @buffer_config.retry_forever
            log.warn "<secondary> with 'retry_forever', only unrecoverable errors are moved to secondary"
          end

          secondary_type = @secondary_config[:@type]
          unless secondary_type
            secondary_type = conf['@type'] # primary plugin type
          end
          secondary_conf = conf.elements(name: 'secondary').first
          @secondary = Plugin.new_output(secondary_type)
          unless @secondary.respond_to?(:acts_as_secondary)
            raise Fluent::ConfigError, "Failed to setup secondary plugin in '#{conf['@type']}'. '#{secondary_type}' plugin in not allowed due to non buffered output"
          end
          @secondary.acts_as_secondary(self)
          @secondary.configure(secondary_conf)
          if (@secondary.class.to_s != "Fluent::Plugin::SecondaryFileOutput") &&

View on GitHub (pinned to dd45c6e18d)

Solutions

  1. Remove the `<secondary>` section from the non-buffered output
  2. Use `@type copy` with two destinations if the goal is dual delivery, not failover
  3. Switch the primary to a buffered output (file, forward, s3, ...) if you actually need retry + secondary semantics

Example fix

# before
<match app.**>
  @type stdout
  <secondary>
    @type secondary_file
    directory /var/log/fluent/backup
  </secondary>
</match>

# after (stdout is synchronous; no secondary is possible)
<match app.**>
  @type stdout
</match>
Defensive patterns

Strategy: validation

Validate before calling

fluentd --dry-run -c /etc/fluent/fluent.conf

Type guard

null

Prevention

When it happens

Trigger: `<match **> @type stdout ... <secondary> ... </secondary>` or `@type copy`/`@type null` with a nested `<secondary>` element; `@secondary_config` is non-nil while `@buffering` is false, so output.rb:401 raises during configure.

Common situations: Pasting a `<secondary>` block into a config whose primary was switched to a synchronous plugin; misunderstanding that `<secondary>` is not a general 'send a copy elsewhere' directive (that is `<match>` + copy).

Understand the failure class

Background: Config validation failed: what "invalid value for {key}" and settings-rejection errors mean across 19 open-source libraries — this error's family across 19 libraries.

Related errors


AI-assisted analysis of fluent/fluentd@dd45c6e18d (2026-08-21). Data as JSON: /api/errors/549393b76a38995a. Report an issue: GitHub.