When using the 1.56 update with Docker, all pollers are being reset back to the general polling group on the first poll.
When Validating my config:
Steps to reproduce:
- Used docker image to build the first poller
- Changed server name in “Specific URL” (base_url) setting
- Enabled distributed polling in global config and set the memcached settings
- Then added a device, waited for pollers to bring in and display data successfully
- Added a second polling group
- Used docker image to build a second poller.
- Added another device to second group
- Waited for poller to show up in the web gui
- Changed the poller to poll only the second group
- Poller ran once
- Checking the poller group settings shows that the second poller has been changed back to being in the first poller group and removed from polling the second group
How did you accomplish Step 9 ?
poller groups are only changed at restart of the dispatch service (this happens during maintenance daily or manually)
Would be possible to make changes applied live through redis, but no one has contributed code for that. (hint probably not a simple task)
This is done through the gui on 1.56 under poller/settings
So if I want to move a poller to a different poller group, I have to restart the dispatch service?
yes, but you should not be doing that on a regular basis. May I ask why you would need to do that more than never or once? Maybe I can help you find a better solution.
Well, we want to be able to put our pollers into different groups. This is something that essentially we will only have to do once. When we change them to a different poller group though, They will change momentarily into the desired group, and then go back to the default group.
There seems to be something a bit funky going on right now.
One user reporting restarting the service immediately after apply the groups made them persistent.
