We came across a bit of an issue when deploying all sitecore marketing definitions, during a recent upgrading our sitecore instance from 9.3.0 to 10.1.2. The deploy marketing definitions tasks appeared to go on for more than 72 hours. This slowed down our upgrade process. So I thought it worth mentioning, on the off chance it helps someone else in the same situation.
I had successfully done two full upgrades of our test environments. During each I documented each of the steps involved and roughly how long they took. This hopefully meant that when it came to doing the upgrade of production, there would be no hidden suprises.
The issue
Most people will know that when it comes to doing a sitecore upgrade, there are many steps that you need to follow. It is important to follow the steps in the guide correctly and in order (or bad things happen - see this post where we lost all out data as a result of not doing this!).
So it was that I was following the steps and I got to the stage of "Deploying Marketing Definitions". Normally, you select all checkboxes and then hit deploy, the spinner does it thing for a while then says the process is complete. This often isnt the case though, as if you look in the jobs viewer the jobs get queued up and sometimes run for some time after.
according to the jobs viewer was still running 72 hours later
Our issue was that the tasks started, but according to the jobs viewer was still running 72 hours later. This posed a problem, as it occured during the upgrade of our main production platform. So, whilst most of the upgrade was completed and the site was up and running. We couldnt offcically close off the process, as we had a number of post upgrade tasks still remaining to do. because of this we could not allow all the user to start doing there thing.
The solution
We didnt want to start killing off the task by recycling the instance. Just in case it caused further problems later down the line. So we contacted the team at sitecore support for some clarification. There analysis of the logs showed many attempts to perform the deployment. But importantly there were no logs showing the status of the newly deployed items (as part of the task). Nor were there any to indicate the deploymeny had finished or failed... there conclusion was that the process must have gotten interupted somehow, prior to fully starting.
To confirm this we opened the /sitecore/admin/jobs.aspx page, set "Refresh each 1 sec". If we saw a constant stream of publishing jobs in starting and stopping, then that would mean the task was still under way. There was no such evidence, so we were able to confirm that the task was not running.
With this confirmed, we were able to recycle the instance to clear the stalled jobs. Then kick off the new redeployment tasks. Sure enough, when we viewed the jobs.aspx page, we saw a constant stream of publishing jobs. Luckily this time though, the task only lasted one hour. So we were able to move on and finialise the deployment.