Purge the pipes

For reasons unexplainable I have occasionally found an alarm that is ‘stuck’ in a pipeline.

Any way to purge the pipes? (apart from resetting the gateway)

This could really be handy if things went south for those a little less clever with looping logic as well I am guessing…

Hi,

First off, we should probably make sure that it’s not a problem with one of the pipeline blocks. Can you maybe post a screen shot of the Status>Alarm Pipelines screen when it happens? What is the final state that it sits in?

Some way to clear the pipeline would be good, and should probably show up on that screen if you’re logged in. When you save the pipeline, the previously running version gets retired, but I don’t think that that drops the existing events out…

Finally, for what it’s worth, there actually is a built in limit to prevent infinite loops- alarm events can only transition between block 1024 times (although should anyone ever need more than that, it’s possible to reset the counter through the Set Property block)

Regards,

How can you tell it is in the pipeline? Is there somewhere you can monitor it?

You can monitor the pipelines from Status>Alarm Pipelines in the gateway. It will show everything currently in them, and which block their in.

Regards,

I’ve got some hanging out there as well in the transition state. Been there for 24 hrs.


If it’s still in that state can you grab a thread dump and attach it to this post?

Here is the thread dump - hope it helps.
thread-dump.txt (148 KB)

Hi,
I have the same problem:
On Notification active pipeline it work ! but on Clear pipeline it doesn’t work.
A function to purge would be cool :thumb_right:
Thanks
thread-dump.txt (37.5 KB)


We’ve both fixed some lingering issues with events getting ‘stuck’ AND added a cancel link next to each event in the status area. :thumb_left:

Good job :thumb_left:

I’m still have a problem with “stuck” tags in the alarm pipeline. I’m running Ver. 7.9.5 and was working with support by email for over 4 months. The problem is still in my project! I have emailed over 3 weeks ago about this and no one has reply this last time. Any suggestion? My project has one data type tag and 255 data type instance tags of it, totaling about 610 tags per data type instance. Plus about 150 regular tags. About 25% are alarm tags and 30% are history tags. So the grand total of tags is roughly 155,000 tags and I wonder if I need to do some kind of data type instance tag scan to make sure all of the data from my data type instance is correct in my instances. Any suggestion because I stuck know!

Thanks,