How to purge/clean-up a Dev/Test environment? We have had an on-going

How to purge/clean-up a Dev/Test environment?

We have had an on-going issue in our development environment that seems to have originated when it was not shutdown properly and KDBs got corrupt. We experience issues where incorrect sub processes are sometimes triggered, or the system says that a process model has not been found. When we try the same request again it works.

What we would like to do is purge all data in the environment, apart from the process models and user accounts. Can you please advise how we should go about doing this? We have tried deleting every single process through the Appian UI, but the problem remains. We also tried rebuilding analytics from scratch but that also doesn't help. Checkengine shows everything as normal.

thanks a lot.

OriginalPostID-174582

OriginalPostID-174582

  Discussion posts and replies are publicly visible

  • I don't think this is related to the improper shutdown. What's the exact error, is it something like: "Process model not found on Execution server for sub process invocation"
  • Hi Eduardo. Sorry for the delay on this. The environment is still not behaving correctly (its not too bad apart from some process instances not starting as expected). I can now confirm one of the errors to you - when we try to start a process in debug mode (via process modeller) it sometimes doesn't start and we see an error "Process Model Not Found". after republishing and trying again it works.

    We have most probably made things worse by trying to "fix" the issue ourselves by deleting and copying KDBs. What I did in my last attempt was to delete/clear all analytics KDBS and I also deleted the KDBS in 2 of the exec engines and copied the KDB from the other exec engine so all 3 engines referenced the same copy of the KDB.

    Just to confirm that what we ideally want to do is start with a clean slate (i.e. purge all process data) and only retain our process models in this environment. Is there a way to do this? thanks
  • 1. If the error you were encountering was: "Process Model not Found on Execution server for sub process invocation" the solution was to individually launch one instance of the problematic model in standalone mode instead of as a sub-process to force the propagation between DESIGN and EXEC. See my comment at forum.appian.com/.../e-152404

    2. You should not copy the pe*.kdb file from EXECX to EXECY. That puts the system in an inconsistent state. Each EXEC has an ID registered inside and knows whether it is 00, 01 or 02. If you copy the pe*.kdb from EXEC00 to process/exec/01/gw1 then you're duplicating an engine causing inconsistency because Appian now has two engines with the same ID and one is the wrong location.

    Instead what you can do is:

    1. Stop the environment
    2. Blank out analytics
    3. Blank all the execs, basically their gw1 directory (since you need to purge and fix the inconsistency)
    4. Start the environment
  • Will deleting the exec engines just delete all process instances or does it have any other side effects?
  • I should have mentioned that this is by no means a standard practice you should be executing on your own. Appian Technical Support will assist you further with the steps.