I've curious/funny question on you who ever met with situation when Auto-start process sometimes not started....
It's need to be pointed that I've correct start condition and also rights to run it (admin). Problem is that only "solution" for that is restart Clarity services on server.
Then everything is working fine, auto-start working as expected...this lasts one day. Another day if I'd like to test it, it doesn't work. So again I've to restart all services, then
it's working whole day. Any tips/ideas? I'm quite confused of this behaviour.
Process Background:Our process is started when user change some custom attribute (DG stage status) in project from one value to another. Then it should be triggered.
Everything works after restarted services. Not before. Condition is correct and following:
If you need any other info, please let me know. Many thanks for your hitns.
Have you checked these ?
Clarity process start with delay of more than a min
RE: Process Start Delay
Unfortunately there is no any answer on my issue...
Hi Matej. What if the solution was to not try to figure out why your process engine is letting you down (problem management), but engineer your system in a way that helps it not let you down (eliminate incidents)?
- Your problem above reads as if you're only running one process engine bg service. What if you add a second service on a different server? This minimizes the risk if one goes down.
- Do you do any scripted starts and restarts? We recycle all services once a week and call a 'start all' nightly. gcubed has a good idea here. Why not call a start every X minutes? It don't cost nuthin' if everything is up and running. When the problem occurs, do you observe your bg process down or is it just unresponsive? If down, definitely look at calling a start on that service every 15 minutes (or any frequency you wish) and see if this makes these incidents just go away.
- Are you on premise or on demand? For super mission critical processes on premise I've gone as far as creating what I call 'backstop' stored procedures to check if that thing that was supposed to happen happened and if not to just take care of it. Just looking at your screen shot above, one could add the Project DG Stage Status to the audit trail, then create a job that every 5 minutes check all projects that had an update and a change in the last X minutes from a user not admin then look at the audit trail and see if there was a change then check your next expected state (be it the running process or any expected results of that process). If the thing that the process engine was supposed to make happen either didn't happen or wasn't happening - kick off whatever you need to do to clean it up. Yes, that this is alot of work - but a pound of prevention saves tons of grief.
I agree I wish everything just worked - but until it does, we have to jump through all the hoops to make it work. To my customers, it always appears as if everything just worked - and that makes my world better.
Nice post, but I have some notes there...
- Yes, in fact, we are running only one bg process in spite of we have 2 Clarity instances. 2 BG processes were causing a lot of problems, which even CA couldn't resolve so only "solution" was shutdown the second BG, so only one remained....
- Restarting All services is quite a rough "solution" for me but I know it may help. But still not solve an issue, it's just usable workaround...
- This process or situation can be handled of course in many ways...as you mentioned. However auto-start process is the simplest and the most elegant solution I'd say which should work perfectly in an ideal world
If any ideas/suggestions will fail, I'll have to handle it somehow by your ideas above - I know all these, but I consider it as still workaround for my problem
In spite of that, thanks for your advices
I have same problem in production environment. (Version 14.1) Sometimes processes do not start or get error without any reason.
I opened a case to the support and here what they said to me.
""Based on the symptoms you are seeing with the processes, you are currently hitting the following bug:
CLRT-67832: UserSecurityPermissionRoleCache is getting cleaned during process execution resulting in "not authorized" errors
Most customers will see this problem very intermittently. Restarting the bg(s) and retrying the problem steps will resolve the problem temporarily.
Customers will see process errors like:
Note this defect is resolved within Clarity 14.2 We strongly recommend you to upgrade to Clarity 14.2 as the Clarity 14.1 release will not have patches, service packs or further development.""
Now we are going to upgrade to version 14.3 in couple weeks. I will check this issue after upgrade would write back my results.
thanks for sharing this . I'm not referring to any kind of errors you described, but it furthermore seems like a bug with BG process when only restart helps...
So only workaround will be applied as upgrade is very far-away future... Thanks again.
I face this issue in 14.3 .. Raised a ticket. Advice from CA Support was to upgrade to 14.4 since apparently the issue is fixed in 14.4 .. Any one else in same situation ?
We also have the same issue - sometimes the auto-process fails giving the message "You do not have the rights to perform the action" even if the user has all required rights to start the process.
It could be resultant of a bug CLRT-67832 UserSecurityPermissionRoleCache is getting cleaned during process execution resulting in "not authorized" errors where caches are not getting updated properly causes the rights to get out of sync. May I know which PPM version you are facing.
Suman we are on v14.2
We just upgraded to version 15.2 and still see this issue. Is bug CLRT-67832 fixed yet? we received below error