Manager Metadata

Manager nodes contain an XML metadata stream called info, which you can use to monitor and set Backburner Manager behaviour related to job processing. It can be obtained using the getMetaData method on the WiretapNodeHandle object for the manager node, or using the wiretap_get_metadata. It is set using the setMetaData method and wiretap_set_metadata command.

The returned metadata is structured as follows:

<info>
  <administrators>root,admin1,admin2,...,adminn</administrators>
  <config>
    <retryCount>server_retries</retryCount>
    <timeBetweenRetriesMS>retry_delay</timeBetweenRetriesMS>
    <maxConcurrentJobs>jobs_on_network</maxConcurrentJobs>
    <forceSingleTaskProcessing>bool</forceSingleTaskProcessing>
    <onJobCompletion>action</onJobCompletion>
    <archiveDays>num_days</archiveDays>
    <deleteDays>num_days</deleteDays>
    <mailServer>smtp_server</mailServer>
    <logLevel>level</logLevel>
  </config>
</info>

The following table explains the elements in the XML stream.

Element	Type	Comments
<info>	complex	Container for the other elements.
<administrators>	string	A comma-separated list of users with admin privileges. Read-only.
<config>	complex	Container for elements to configure the behaviour of the Backburner Manager.
<retryCount>	integer	Number of times the Backburner Manager attempts to restart a job on a server that has failed to complete its processing. A failed job might be returned by Backburner to the job processing queue. Set to zero (0) to have job processing halted on the server after its first failure.
<timeBetweenRetriesMS>	integer	The time (in milliseconds) before the Backburner Manager attempts to re-start a job on a server that has failed. Works in conjunction with <retryCount>. Default is 30,000 milliseconds (30 seconds).
<maxConcurrentJobs>	integer	Maximum number of jobs that Backburner sends out for processing on the render farm at the same time.
<forceSingleTaskProcessing>	boolean	By default, the Backburner Manager can assign multiple tasks or blocks of tasks to a single server. To force the Backburner Manager to assign each server with only one task at a time, set this boolean to 1 (true).
<onJobCompletion>	string	A token specifying what happens to job metadata once the job has successfully completed. Works in conjunction with <archiveDays> and <deleteDays>. Valid values: leave: Leave in the job list. delete: Remove from the job list after the number of days specified by the <deleteDays> element. archive: Remove from the job list and place in the job archive after the number of days specified by the <archiveDays> element.
<archiveDays>	integer	Number of days a completed job is left in the job list before being archived. Set to 0 (zero) to archive the job immediately after its completion.
<deleteDays>	integer	Number of days a completed job is left in the job list before being deleted. Set to 0 (zero) to delete the job immediately after its completion.
<mailServer>	string	The Backburner Manager can send job success/failure notifications to the addresses submitted with the job (see Job Metadata). This element indicates the server where the smtp mailer daemon is running.
<logLevel>	string	Determines the level of information collected and written to the log file maintained by the Backburner Manager (backburner.log). Valid values: error: Records fatal operation failures. warning: Operations that complete with non-fatal errors. info: Successful operations, possibly with minor faults or caveats. debug: Detailed state information includng TCP/IP packet information that is helpful in tracking down bugs. debugX: Extended debug information. For more information, see the Backburner User Guide.

Example

The following use of the wiretap_get_metadata command returns the metadata for the Backburner manager (always named /) on a host called rossendale:Backburner.

In this case, the Backburner Manager is set to retry a failed job on the same server three times, waiting 30 seconds before reinitializing the job. The Backburner Manager is configured to issue a maximum of 20 jobs onto the render farm at once, and to archive completed jobs after five days. Any mail notifications are sent out through a daemon running on host copenhagen. The log level is error.

wiretap_get_metadata -h rossendale:Backburner -n / -s info
<info>
  <administrators>root,m_flinders</administrators>
  <config>
    <retryCount>3</retryCount>
    <timeBetweenRetriesMS>30000</timeBetweenRetriesMS>
    <maxConcurrentJobs>20</maxConcurrentJobs>
    <forceSingleTaskProcessing>0</forceSingleTaskProcessing>
    <onJobCompletion>archive</onJobCompletion>
    <archiveDays>5</archiveDays>
    <deleteDays>5</deleteDays>
    <mailServer>copenhagen</mailServer>
    <logLevel>error</logLevel>
  </config>
</info>