We went live with 7.5.0.1 in Production last night (on new linux servers) and one of the tasks in our process was to do a final copy of the backing store from the old server to the new server. We discovered that the existing workflows were unable to continue on to their next tasks after they were copied to the new server. The reason is because we use this syntax to define the command for externaltasks in WFM:
http://$IW_SERVER/iw-cc/urlexternaltask
which embeds the actual hostname into the command. After the workflows ended up on a new server, the command could not resolve any longer and those tasks got stuck.
I'm wondering if it's safe, and more importantly, will that problem be resolved if we use this syntax instead:
http://localhost/iw-cc/urlexternaltask
We have a failover server and a disaster recovery server in our prod environment, and if we ever had to move to one of those other servers, we'd run into the same issue. The workflows would all be dead in the water. If we use localhost in the URL for command, will it just naturally work from server to server, regardless of the state of the workflow when it is running on a new server?
Anyone have a better solution?
