ML 8中如何通过代码检测任务服务器任务完成状态？

Absolutely, you can fully automate your batch workflow in MarkLogic 8—no more manual checking of the task server queue! Here's a step-by-step solution to detect when tasks finish and chain your dependent batches programmatically:

1. First, Create a Helper Function to Check Queue Status #

You can query the task server's status to get the current queue size using MarkLogic's built-in admin functions. This function will return true when the queue is empty:

import module namespace task = "http://marklogic.com/xdmp/task" at "/MarkLogic/task.xqy";

declare function local:is-task-queue-empty() as xs:boolean {
  (: Get the first task server on your host (adjust if you have multiple) :)
  let $task-server := xdmp:servers()[xdmp:server-kind(.) = "taskserver"][1]
  let $server-status := xdmp:server-status(xdmp:server-id($task-server))
  (: Extract the current queue size from the status :)
  let $queue-size := $server-status/server-status/queue-size
  return $queue-size = 0
};

2. Add a Wait Function to Poll Until Queue is Empty #

To avoid spamming the server, we'll add a delay between checks. This function will loop until the queue clears, with optional logging to track progress:

declare function local:wait-for-empty-queue($wait-seconds as xs:integer := 10) {
  while (not(local:is-task-queue-empty())) {
    (: Wait for the specified interval :)
    xdmp:sleep(xs:dayTimeDuration("PT" || $wait-seconds || "S"))
    (: Log current queue size for visibility :)
    let $current-size := xdmp:server-status(xdmp:server-id(xdmp:servers()[xdmp:server-kind(.) = "taskserver"][1]))/server-status/queue-size
    xdmp:log("Waiting for task queue to empty... Current queue size: " || $current-size)
  }
  xdmp:log("✅ Task queue is empty! Proceeding to next batch.")
};

3. Chain Your Batch Tasks #

Now you can structure your workflow to submit a batch, wait for it to finish, then submit the next dependent batch. Here's a complete example:

(: Define your task functions first :)
declare function local:generate-xml-file($output-dir as xs:string, $task-id as xs:integer) {
  (: Generate your XML content here :)
  let $xml-content := <document><id>{$task-id}</id><content>Generated from first batch</content></document>
  (: Save to D drive (note Windows path escaping with double backslashes) :)
  let $file-path := $output-dir || "batch1_doc_" || $task-id || ".xml"
  return xdmp:save($file-path, $xml-content)
};

declare function local:generate-dependent-xml-file($output-dir as xs:string, $task-id as xs:integer) {
  (: Access files from the first batch here (adjust logic as needed) :)
  let $source-file := $output-dir || "batch1_doc_" || ($task-id mod 10000) || ".xml"
  let $source-content := if (fn:doc-available($source-file)) then fn:doc($source-file) else <default/>
  (: Generate dependent XML content :)
  let $xml-content := <dependent-document><id>{$task-id}</id><source>{$source-content}</source></dependent-document>
  let $file-path := $output-dir || "batch2_doc_" || $task-id || ".xml"
  return xdmp:save($file-path, $xml-content)
};

(: --- Start the workflow --- :)
(: Submit first batch of 10,000 tasks :)
xdmp:log("🚀 Submitting first batch of 10,000 tasks...")
for $i in 1 to 10000
return task:add-task(
  xdmp:function(xs:QName("local:generate-xml-file")),
  ("D:\\output\\", $i),
  <options xmlns="http://marklogic.com/xdmp/task">
    <priority>normal</priority>
  </options>
);

(: Wait for first batch to complete :)
local:wait-for-empty-queue(10);

(: Submit second batch of 20,000 tasks :)
xdmp:log("🚀 Submitting second batch of 20,000 tasks...")
for $i in 1 to 20000
return task:add-task(
  xdmp:function(xs:QName("local:generate-dependent-xml-file")),
  ("D:\\output\\", $i),
  <options xmlns="http://marklogic.com/xdmp/task">
    <priority>normal</priority>
  </options>
);

(: Wait again before next batch :)
local:wait-for-empty-queue(10);

(: Add subsequent batches here following the same pattern :)

Key Notes & Considerations #

• Permissions: Make sure the user running this code has the task-admin role (or at least permissions to read server status and add tasks).
• Task Server Selection: If you have multiple task servers, adjust the $task-server selection logic to target the correct one.
• Wait Interval: Tweak the $wait-seconds parameter based on how fast your tasks run—shorter intervals for fast tasks, longer ones for slower jobs.
• Error Handling: You might want to add error handling (e.g., checking if files exist before processing) to make the workflow more robust.
• Logging: Check MarkLogic's server logs (usually in C:\Program Files\MarkLogic\Data\Logs on Windows) to track the workflow's progress.

内容的提问来源于stack exchange，提问作者Dixit Singla