Scheduled Agent stops running after one week

HCL Domino Domino Forum
1

I have an agent that is part of a database that ceases to run after 1 week exactly. I need to restart the server for it to be picked up. The agent is a LotusScript agent that runs ever 10 minutes. Run on All documents in database. It picks up a file on the hard drive, in a folder with access to everyone to read/write. It reads and kills the file. It then writes to another file in the “Users\Administrator” folder. Things I’ve tried include:

  1. Checking the script libraries so they are signed properly.

  2. Database is allowing background agents.

  3. Agent is signed with Admin id. Full rights.

  4. Security level is set to “3. Allow restricted…”

  5. Split agent up into multiples, running on different days, same effect.

  6. Put agent into fresh db, same effect.

  7. Put agent into db where agents always work fine, same effect.

Like I said, it runs for one week and even though I’ve tried everything to get it to be added back to the scheduled agents, it won’t. It gets bumped from the scheduled agents and only a server reboot will bring it back. I have other agents that do the same type of thing, run very frequently, but they don’t interact with the file system. I thought maybe it was having rights issues writing to the folders, but how can that rationalize it being kicked from the list of scheduled agents? And after one week exactly.

Any thoughts are much appreciated. Thank you for reading.

2

Subject: RE:

Not sure if this is the case but worth checking anyway:

Scheduled agent stops executing if Compact is already running on the database
http://www-01.ibm.com/support/docview.wss?uid=swg21084223 http://www-01.ibm.com/support/docview.wss?uid=swg21084223

3

Subject: question

Did anything strange happen on the last run of this agent? Like file system being backed up? Could it be max concurrent agents are kicking off at the same time? I can suggest having a program doc run on the weekend to run “tell amgr sched” Here is a technote on how to configure program docs to run console commands http://www-01.ibm.com/support/docview.wss?uid=swg21088932 http://www-01.ibm.com/support/docview.wss?uid=swg21088932

4

Subject: Does another scheduled agent run on this database?

By default only one scheduled agent runs at a time on a database.

What do the server’s Agent Managers say about what’s running on this database?

Running agents every 10 minutes is bound to cause issues, but they don’t cause this particular issue. Eventually though your admins will be annoyed. The schedule queue becomes useless with the agent queued up like this.

If it’s just doing file operations, I’d probably suggest pushing it out to VBScript and schedule it in Windows. Because it doesn’t seem to offer much use in Domino, based on its description.

People have replied with some good, plausible causes.

I’ve run into shared drive issues where due to network issues the OS will stop responding, freezing the agent. Net result: the agent stops running.

There can be lock-ups with multiple agents racing for the same file. E.g., a test & prod agent both looking for the same file, then one deletes it out from under the other. Or a separate agent running infrequently that locks up the file.

Does your server do design updates? An update could change the signatures on script libraries or it may have a later signature on the agent. Script libraries also have to be signed with sufficient access rights.

It may be worthwhile looking at the agent manager’s queue to see if something else is gumming up the schedule queue. Is this agent truly falling off the queue?

Is the server or agent manager telling you it’s falling behind? If it is, well, it is. In that case you’d need to reduce the runtime of the agent, because it’s taking longer to run than the 10-minute interval given to it. e.g., 10 minutes when you’re letting agents run for 15 minutes, means you could lap the server.

5

Subject: Answer

Make sure that the correct number of agent managers are running on your system and that the timeouts allow enough time for the agent to run.

tell amgr status

02/15/2016 09:41:09 AM There are currently ‘2’ Agent Executives running
02/15/2016 09:41:09 AM There are currently ‘27’ agents in the Scheduled Task Queue
02/15/2016 09:41:09 AM There are currently ‘0’ agents in the Eligible Queue
02/15/2016 09:41:09 AM There are currently ‘11’ databases containing agents triggered by new mail
02/15/2016 09:41:09 AM There are currently ‘12’ agents in the New Mail Event Queue
02/15/2016 09:41:09 AM There are currently ‘1’ databases containing agents triggered by document updates
02/15/2016 09:41:09 AM There are currently ‘1’ agents in the Document Update Event Queue

  1. Check database consistency and ods version

  2. Deletion stubs, if your deleting allot of content you will need to do database maintenance.

6

Subject: Enable agent manager debug

If you haven’t done so already, please enable agent manager debug:

tell amgr debug *

Then see what the logs show once the issue returns. I would also check to see the last modified in the designer client (Code → Agents) to see if it was modified in any way by an id file after it had started.

It might also help to create a program document that outputs - “tell amgr schedule” every hour or so. That way you can find out when it stops showing up. That will help you review the console.log file to see what was going on in the server around that time.

7

Subject: Maybe something like this?

Just an example
Mapped drive is disconnected - Windows Client | Microsoft Learn https://support.microsoft.com/en-us/kb/297684