Agents Agents Everywhere

Agents standing for Scheduled agents are the most commonly misunderstood things in livelink .I have senior livlink people, friends,who think that the Admin Service is responsible for running schedulers.In any case since I started early in version 8.1.5 there was a 3rd service called ‘Change Agents’. I digress but here’s what I wanted to write.

If you see in your opentext.ini a line called loader=sockserv;agents;notify     these are insructions to the LLserver to run those as “threads” in the LL System.For almost all livelink history agents and notify are run in separate threads.Notify runs in the notify thread for the list it controls and agents in its list.


See official OT Links

The above link was published by OT when I as a junior admin(2003/2004 period) tried to debug the extreme load on our system my senior colleague John Simon and Mark Simm were on email and phone and both were very good on oscript.I slowly learned oscript to kind of become good at it.Right now very few companies split notifications and the current OT code has bugs which I have reported.But it will affect companies who are doing split notifications as nobody usually does it.

I have a problematic agent.It is common knowledge or with a little bit of research one can figure out how agents are run.For the record here’s what it would take to make an agent id 0f 222378 I made that number up run in a LIVELINK server.

  1. Create module and write the agent code.Put 222378 as its id.Register the agent with the livelink tables either using SQL or a weblingo file.If everything goes well you can now run a query like this to find it select * from agentschedule where agentid=222378
  2. It is quite possible that in your opentext.ini it will put under this area something like this
  3. [scheduleactivity]
    222378=1 //this means run THIS agentid on THIS SERVER
  4. 222379=0 //this means run DO NOT RUN THIS agentid on THIS SERVER
  5. And this line should allow agent threads to run
  6. [loader] //this is the magic line for making a livelink server a agent/notify server
    load_relagent=relagent //Notice there are so many ways to split agents
  7. In small livelink organizations the Admin server is used to run notifications,agents and wfagents that is the reason why everybody thinks admin service is responsible for something.It is not faintly connected to scheduling.Usually that is the only server that would have that line like this

load=sockserv;javaserver;agents;notify;wfagent                               most front end servers will say load=sockserv;javaserver;

This is needed because otherwise all the agents will start at the same time.You can avoid it by putting the agentid=0 in other servers.

Here’s what will happen at about 5 minutes since a livelinkserver is started not based on the computer clock,this agent thread starts up as agents101.out.The first thing it will do is run this oscript line $LLAgent.AgentController.New().Execute( .fPrgCtx, args.Agent_list, args.ExcludeAgent_list ) which means run one of those controller lines with this parameters.So to run it for debugging you can do something like

?func=agent.runagent&agent_list={‘222378’} so almost like saying hey livelink can you run the code in agentid 222378.

Here’s where I went wrong.say for e.g 222378 was not registered in the DB then this call translates to while I have very intention of running your agent I am not getting a hit for 222378 so the net effect is livelink will run all outstanding agents which you are trying not to do.Since agents are serial it will be sometime for the real interested agent to get that call.

To do manual running you usually do this

To test with builder…


– In opentext.ini set/add the following:

– In the browser type the following URL:

When this call hits livelink that thread finds out from the agentschedule table all agents who has the lastruntime in the past.Then it verifies who all are qualified (0 or 1 or no info is regarded as enabled agent ) to run then it prunes the excluded_agent_list and tries to run the agent_list

If one of those agents in the list do not come out clean as in a bug /exception/trace then it will error out and the nexttime is not updated.So it will start a vicious cycle that is when you see like why is my DirSvcs agent who is supposed to run once every day run every five minutes

This is a very common occurence of low agentid’s DirSvcs,OI etc who are in the 2000 range


7 thoughts on “Agents Agents Everywhere

  1. Appu, Good article to understand about the way Agents works.

    We are using manual import option for importing the docs from shared drive and the problem here is the file path. It is exceeding beyond the 256 characters, Is there any way we can change this setting so that I can accept the files.



    1. If you are basically sourcing files using a UNC path and running into limitations mount the directory as a drive that would make your path a little shorter.You could also look at dropping the files to the livelink server and removing them as needed.As such this is not a livelink limitation but something you have to understand and work around.Though not sure I am pretty positive that modern OS’s can handle stuff.If you are sure it is a livelink error you should look at the XML declaration the UTF-8 is more prevalent in my opinion.

  2. We have already added the shared as a drive and using this location we are doing the Import. The other option you mentioned about the XML declaration, you mean to include a tag ”
    “? If we add this then it will help us in reading the large file path (>256 char)??

  3. We have already added the shared as a drive and using this location we are doing the Import. The other option you mentioned about the XML declaration, you mean to include a tag ” “? If we add this then it will help us in reading the large file path (>256 char)??

  4. Can run object importer in all the instances? can enable multiple thread for object importer picking files from multiple location at the same time?

    1. yes it can be done but it is more complicated than you actually think
      That is what we all do the OI from OT is capable of 2990-2999 agentIDS
      so on 2990 we put files for one customer
      so on 2991 another customer
      you have to level set expectations

      for example consider this scenario you say in a control file to upload to a path
      Ent:1:2:3:4:5:myfile.pdf but in actuality Folder 4=3 and 4 and 5 was in another OI control file that may run by agent 2991 this can pose problems unless you had missing container creation ticked. Hence we take some effort in configuring the “method for the madness

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.