Scheduler configuration for clustered Environment

, ,
1

Hi,

We have a application running in HA. here we have two IS pointing to the same database. DB is clustered.
i have few schedular tasks for executing a service, these tasks are configured on both the IS. Now with this setup both the IS runs the schedular task and the service got executed twice. Is there any way to configure the IS, so that if one IS run the schedular task the other IS do not execute their task. or if there is some other way to do it?

Any help will be highly appreciable.

Thx

2

You’ll probably want to use IS clustering if you want the scheduled tasks to be cluster aware (e.g. only 1 runs at a time).

3

Hi reamon,

i dont want to use IS clustering, i want to know if there is any way to execute the scheduled service only once, even if its scheduled to run on both the IS.

4

If you don’t want to use IS clustering then you’ll need to write some helper services to do this.

An approach I’ve used in the past is:

  1. Create a config file (XML) that lists the host names and ports of the IS instances in the order you want.

  2. Create a service (maybe named “should_I_Run”) that determines the name/port of the IS instance it is running on. Then loads the XML file with the list of host names. If the IS instance name matches the first name of the list, then return “true”. If it doesn’t match, ping that IS to see if it responds. If it does, return “false”. If it doesn’t, move to the next name on the list and repeat.

  3. In the service to be invoked by the scheduler call the “should_I_Run” service. If it returns true, continue. If false, stop.

All IS instances would do this same approach. In this way, just one of the IS instances will run the service (typically always the first one in the list). But if the first instance is down, the second will run it and so on.

I used an XML file but the list could be anywhere (text file, DB table, etc.).

HTH

5

That looks a great solution…
thx a lot reamon :slight_smile:

i have one more question,
Presently we have migrated from WM6.5 to WM7.1.2. In wm6.5 the scheduler information was getting stroed in a file called jobs.cnf, now with new setup in wm7.1.2 this file is not present. do u have any idea where the scheduler information is getting store in wm7.1.2, i read in wm advantage that in wm7.1.2 they have moved the scheduler info from the file to db. But in our setup we did not run any migration script to migrate the scheduler data from file to db, and we dont have this jobs.cnf file also in wm7.1.2 environment.
do u have any idea where is these info getting store now?

6

I do not, though I suspect that the migration of the tasks from jobs.cnf to ISInternal happens automatically and does not require specific administrative/configuration action.

7

Hi,#

Schedulers are shifter to IS DB schema where table name starts with Task. This gets updated as per the action performed on the IS GUI on schedulers. Jobs.cnf of 6.X has been removed in the later version and no longer exists.

Regards

Punit kumar

8

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.