If the agent runs the action after restarting the service, it suggests that the Client was not aware of the action when it was issued. This is most likely due to it not receiving the UDP notification which tells it to gather the action/site.
just wondering if i have large set of servers may more than 1000 or more , does it impact to application server or it is safe to use this settings for those client which are not accepting new actions?
While it may be relatively safe, the endpoints that are not able to receive notifications will be less responsive than those that can. It would be good to try to identify why they are not able to receive notifications, and work to address it.
These are windows OS , 2008 , 2008 R2 , local firewall is off at windows level , i think i need to check with network team…
Second query for Linux
i found error status on the linux client on the blank action , but when i restarted agent service on linux host , it start working. Not sure what i should check on linux side , last report time is up to date.
last report time is different than the client processing actions.
the client will send reports periodically no matter what, even if it isn’t getting notified by the relays that there is something new for it to do.
Figuring out the UDP notifications by opening up firewalls is the best bet. The next best thing would be to enable command polling for something like once an hour or so to at least get a maximum amount of time it will take for the client to see new actions.