Skip to main content

Restart a node

A node that stopped reporting, or acts up, often recovers after a restart. Jaah offers two. Restart under More restarts only the node's Jaah service: running sessions keep going, and the node reconnects in under a minute. Restart on a machine Jaah launched in your cloud reboots the whole machine: the node goes offline for a few minutes, and running sessions end. Try the first one first.

Who can do this: both require the org admin role. Without it, More holds no Restart, and the machine's section doesn't show. Restarting every node on a host, or the host itself, is the operator's.

Restart the node's Jaah service​

  1. In the sidebar, select Directory, then the Machines tab.
  2. Select the node's name. Its details open on the Overview tab.
  3. Select More, below the node's fields, then Restart.
  4. Restart node? asks first. Select Restart.
  5. Read the line at the top: Restarting and the node's name.
  6. Wait up to a minute, then check that its dot is green again: see Read a node's status.

Its tasks keep running through the restart. If it fails, the page says Couldn’t restart and the node's name.

Reboot a machine Jaah launched in your cloud​

Only a node Jaah launched in your own cloud account has this. A machine you run yourself is rebooted the way you always reboot it.

  1. Open the node's details, on the Overview tab.
  2. Find the section named for its cloud, for example AWS machine. Its State should read Running; Restart shows only then.
  3. Select Restart.
  4. Restart machine? asks first: the machine reboots, and its disk and type stay. Select Restart.
  5. If work is running there, Restart anyway? opens instead. It names the tasks running, people's open sessions, and any node on the machine that stopped reporting. Select Restart anyway to go ahead, or Cancel to wait.
  6. Wait a few minutes. State shows Running again and the node reconnects by itself.

After Restart anyway, running sessions end. Each task's runner starts again in its slot once the node is back, so no task is lost.

To stop new work before either restart, drain the node first. What each button on the node's details does is in Machines; what a node and its host are, in How nodes work.