MFT Scheduled Events run on only one node in multiple instance environment

, ,
1

What product/components do you use and which version/fix level?

10.7

What are trying to achieve? Please describe in detail.

Issue with MFT Scheduled Events on multiple IS Nodes.

Do you get any error messages? Please provide a full error message screenshot and log file.

Dear Expert, We are seeing this issue with MFT Scheduled Events in multiple IS Nodes(we have 3) environment where in which scheduled Event runs at only one instance at any given point of time ie we aren’t able to set scheduled event to run one its respective indivudal nodes. The moment we change the scheduled event to run on, say instance A, then automatically scheduled event on instance B get also changed to run on instance A . So file dropped from load balancer on Node B is never getting picked up. Any thoughts ?

2

Hi Saurav

When IS is not in cluster, the IS tasks created for MFT scheduled event will be configured as the same IS node as the target node. This is the behavior from MFT end.

User can go to IS admin page and set Target as Any node. IS will take care of executing on any node there after.

Many of our customers are doing this and few have written IS service to update IS task definition using public services exposed in Integration Server.

Hope this helps.

3

Thanks for the response . We are seeing similar behavior with default IS scheduler as well i.e. wm.server.dispatcher.deleteExpireUUID and wm.monitor.admin:imageCleanup too. So, we can see them getting executed to one of the node at a time on both instances. we arent able to make them run on their respective nodes. Any suggestion

4

For HA purposes, you may want to revisit the approach. What happens if node B receives a file, then goes down or is taken down? The file will not be picked up, which you undoubtedly do not want.

One approach:

  • Create/define a shared volume that all instances access.
  • Incoming files get written, by any node, to that shared volume.
  • The receiving node then publishes a document (assuming you’re using Broker or UM) indicating file received.
  • Any node can then process that doc to pick up and process the file as you see fit.

An ancillary benefit is the latency is reduced a bit – don’t have to wait for the scheduled task to run to process the file.

5

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