4

Using WSO2 BPS 3.6.0 we encountered a serious issue

We have a few processes waiting for external events (with timeout) and several processes polling for updates (using wait node).

The problem arises as soon we restart a server: * timeouts which are passed during the downtime are not processed * wait nodes are not processed at all

Reading related articels:

I found that the timeout timestamps are stored in the ode_job table. So I tried to update the timeout timestamps (before starting up the BPS server)

update ode_job set ts=(near_future_timestamp) where ts>(before_restart) and ts<(near_future_timestamp)

which resolved the scope timeouts, however the wait nodes are not processed anymore even they were stated in future. That effectively blocks all the polling instances without any means to move them further.

Is there a way to "revive" or timeout the wait nodes after restarting the server?

gusto2
  • 11,210
  • 2
  • 17
  • 36
  • Missed it earlier. Will see if there is a problem in ODE, but for that I would need a sample process that reproduces the problem and the exact steps. Can you provide it? – Sathwik Apr 09 '18 at 06:11

0 Answers0