-
Notifications
You must be signed in to change notification settings - Fork 11
Checking Querki Cluster State
Justin du Coeur, AKA Mark Waks edited this page Apr 4, 2017
·
1 revision
There are three separate factors in whether Querki is running properly on a given node:
- The node itself must be running.
- ConductR must be running, and connected to the rest of the cluster.
- Querki must be up and running, and connected to the rest of the cluster.
This page is mainly concerned with the last of these. When rebooting a node or restarting ConductR, you should check whether Querki is properly connected before doing a second node. (Querki requires at least two nodes to be running and connected, or it will go down completely.)
Only a Querki Administrator can check this. Note that this is a permission on the Querki User, entirely separate from the operating-system permissions.
- First, log into Querki using an account with Admin permissions.
- There should be an "Admin" menu in the top bar. From that, select "Show Space Status".
- This page shows what's currently going on in Querki: the connected Nodes, which node is the "Leader" (this will generally have slightly higher load than the others), and which Spaces are currently active.
- That page includes a list of "Members". There should be three Members, whose IP addresses match those of the Nodes. All three should show the status
QUp
.
Assuming all three nodes are in QUp, then you're ready to roll: the network is fully connected, and it should be safe to reboot a single node or ConductR. Wait until all three are in QUp before proceeding to the next.