Is there any guidance or recommended frequency for server reboots?
Requirement: It is desired to schedule a regular restart of the Orchestrator nodes and 3 HAA Linux nodes.
Resolution: This is interim guidance pending an official update to UiPath documentation. UiPath does not require or recommend regular reboots unless otherwise specified in the documentation.
Orchestrator does not benefit from regular reboots. Windows Server is robust enough to go years without reboots, especially if SQL Server is running on it.
For the Orchestrator web servers in a cluster, take them offline for rebooting one at a time with a minimum of one always up and running and healthy. Reboot up to two of three at a time, if there is urgency in the procedure.
HAA does not need reboots; it would only be required when updating the system kernel or if required to perform other maintenance. But this is not a UiPath or HAA requirement.
The HAA cluster is a bit more complicated, as you always need 1 more than half of the cluster's servers online to keep it functional. (3 servers - 2 always up, 5 servers - 3 always up.) The way to achieve this is to put one server at a time in maintenance mode, perform any updates or other changes, reboot it, then bring it back online and taking it off maintenance mode. Shift to the next server and repeat.
Below is the more detailed round-robin process:
- rladmin node maintenance_mode on demote_node => put in maintenance mode
- rladmin status => check your nodes. They should be ok, green everywhere and this node should have 0/0 shards
- Perform the required on the server and then bring it back up
- rladmin status => check your nodes. They should be ok, still green and the one in maintenance mode still with 0/0 shards
- rladmin node maintenance_mode off => take out of maintenance mode
- rladmin status => check the nodes. They should be ok, all green with 2 with 1/100 shards and 1 with 0/100 shards
- Then do the same with the other nodes in turn.
UiPath does not require or recommend regular reboots unless otherwise specified in our documentation. At the end of the day, it is recommended to follow local IT team's and Microsoft's requirements and recommendations on whether and when to reboot the systems.