Node Joined With No Data

When a Scylla node joins the ring, it is visible by the other nodes in the cluster in the UJ (Up Joining) state. Its data steadily grows until it has its share of the data and then switches to the UN (Up Normal) state. This troubleshooting guide describes what to do in cases where the node joins the cluster directly into the UN state - without any data.

Note

This procedure only covers the situation in which a new node joins the cluster. Nodes that are replacing existing nodes following the Replace A Dead Node procedure are expected to be in UN state from the start.

The Problem

There are two possible causes for a new node not acquiring data from its peers:

  • The new node’s configuration file (scylla.yml) lists only itself as the seed node and doesn’t use the same seed address as the other nodes in the cluster do.
  • The node was added with the auto_bootstrap flag set to False

Both scenarios constitute user error and are the incorrect way of adding new nodes to the cluster. See Adding New Node for more details. In this article, we will discuss what can be done to safely restore your cluster in case this situation happens.

Warning

It is important that you do not run nodetool cleanup in any of the other nodes in the cluster until this is fixed. Otherwise, data may have already been deleted from the existing nodes.

Solution

Steps to fix:

  1. For every node that joined the cluster directly in UN state, run nodetool decommission serially. Do not decommission more than one node at the same time.

  2. Make sure to clean the data from the decommissioned node

    sudo rm -rf /var/lib/scylla/data
sudo find /var/lib/scylla/commitlog -type f -delete
sudo find /var/lib/scylla/hints -type f -delete
sudo find /var/lib/scylla/view_hints -type f -delete

    Actual directories might be different than above, check your scylla.yaml to validate.

  3. Once all new nodes are decommissioned, add them again with the correct procedure. Do not add nodes as seeds. Do not add nodes with the auto_bootstrap flag set to False.

Other important considerations:

  • Avoid running nodetool removenode for the new nodes. Running nodetool removenode instead of nodetool decommission will cause the new writes that were sent to the new nodes to be lost. In all cases, you should always prefer decommission over removenode. Use removenode only cases where there is no other way to restart the node.
  • While running a repair in the cluster will bring some data back to the new node and may be a way of restoring its data, it will take much longer than the procedure suggested. It is also important to keep in mind that one copy is lost per each node that was inserted directly into the UN state. In the extreme case in which N nodes were added where N is your replication factor, some data may have been permanently lost and repairs will not be able to restore it. Therefore, repair is not recommended to fix this issue, except in situations where the user is aware and their consistency model can tolerate missing data points.