orama/docs/COMMON_PROBLEMS.md
anonpenguin23 f26676db2c feat: add sandbox command and vault guardian build
- integrate Zig-built vault-guardian into cross-compile process
- add `orama sandbox` for ephemeral Hetzner Cloud clusters
- update docs for `orama node` subcommands and new guides
2026-02-27 15:22:51 +02:00

5.3 KiB

Common Problems & Solutions

Troubleshooting guide for known issues in the Orama Network.


1. Namespace Gateway: "Olric unavailable"

Symptom: ns-<name>.orama-devnet.network/v1/health returns "olric": {"status": "unavailable"}.

Cause: The Olric memberlist gossip between namespace nodes is broken. Olric uses UDP pings for health checks — if those fail, the cluster can't bootstrap and the gateway reports Olric as unavailable.

Check 1: WireGuard packet loss between nodes

SSH into each node and ping the other namespace nodes over WireGuard:

ping -c 10 -W 2 10.0.0.X   # replace with the WG IP of each peer

If you see packet loss over WireGuard but not over the public IP (ping <public-ip>), the WireGuard peer session is corrupted.

Fix — Reset the WireGuard peer on both sides:

# On Node A — replace <pubkey> and <endpoint> with Node B's values
wg set wg0 peer <NodeB-pubkey> remove
wg set wg0 peer <NodeB-pubkey> endpoint <NodeB-public-ip>:51820 allowed-ips <NodeB-wg-ip>/32 persistent-keepalive 25

# On Node B — same but with Node A's values
wg set wg0 peer <NodeA-pubkey> remove
wg set wg0 peer <NodeA-pubkey> endpoint <NodeA-public-ip>:51820 allowed-ips <NodeA-wg-ip>/32 persistent-keepalive 25

Then restart services: sudo orama node restart

You can find peer public keys with wg show wg0.

Check 2: Olric bound to 0.0.0.0 instead of WireGuard IP

Check the Olric config on each node:

cat /opt/orama/.orama/data/namespaces/<name>/configs/olric-*.yaml

If bindAddr is 0.0.0.0, the node will try to bind to IPv6 on dual-stack hosts, breaking memberlist gossip.

Fix: Edit the YAML to use the node's WireGuard IP (run ip addr show wg0 to find it), then restart: sudo orama node restart

This was fixed in code (BindAddr validation in SpawnOlric), so new namespaces won't have this issue.

Check 3: Olric logs show "Failed UDP ping" constantly

journalctl -u orama-namespace-olric@<name>.service --no-pager -n 30

If every UDP ping fails but TCP stream connections succeed, it's the WireGuard packet loss issue (see Check 1).


2. Namespace Gateway: Missing config fields

Symptom: Gateway config YAML is missing global_rqlite_dsn, has olric_timeout: 0s, or olric_servers only lists localhost.

Cause: Before the spawn handler fix, spawnGatewayRemote() didn't send global_rqlite_dsn or olric_timeout to remote nodes.

Fix: Edit the gateway config manually:

vim /opt/orama/.orama/data/namespaces/<name>/configs/gateway-*.yaml

Add/fix:

global_rqlite_dsn: "http://10.0.0.X:10001"
olric_timeout: 30s
olric_servers:
  - "10.0.0.X:10002"
  - "10.0.0.Y:10002"
  - "10.0.0.Z:10002"

Then: sudo orama node restart

This was fixed in code, so new namespaces get the correct config.


3. Namespace not restoring after restart (missing cluster-state.json)

Symptom: After orama node restart, the namespace services don't come back because RestoreLocalClustersFromDisk has no state file.

Check:

ls /opt/orama/.orama/data/namespaces/<name>/cluster-state.json

If the file doesn't exist, the node can't restore the namespace.

Fix: Create the file manually from another node that has it, or reconstruct it. The format is:

{
  "namespace": "<name>",
  "rqlite": { "http_port": 10001, "raft_port": 10000, ... },
  "olric": { "http_port": 10002, "memberlist_port": 10003, ... },
  "gateway": { "http_port": 10004, ... }
}

This was fixed in code — ProvisionCluster now saves state to all nodes (including remote ones via the save-cluster-state spawn action).


4. Namespace gateway processes not restarting after upgrade

Symptom: After orama upgrade --restart or orama node restart, namespace gateway/olric/rqlite services don't start.

Cause: orama node stop disables systemd template services (orama-namespace-gateway@<name>.service). They have PartOf=orama-node.service, but that only propagates restart to enabled services.

Fix: Re-enable the services before restarting:

systemctl enable orama-namespace-rqlite@<name>.service
systemctl enable orama-namespace-olric@<name>.service
systemctl enable orama-namespace-gateway@<name>.service
sudo orama node restart

This was fixed in code — the upgrade orchestrator now re-enables @ services before restarting.


5. SSH commands eating stdin inside heredocs

Symptom: When running a script that SSHes into multiple nodes inside a heredoc (<<'EOS'), only the first SSH command runs — the rest are silently skipped.

Cause: ssh reads from stdin, consuming the rest of the heredoc.

Fix: Add -n flag to all ssh calls inside heredocs:

ssh -n user@host 'command'

scp is not affected (doesn't read stdin).


General Debugging Tips

  • Always use sudo orama node restart instead of raw systemctl commands
  • Namespace data lives at: /opt/orama/.orama/data/namespaces/<name>/
  • Check service logs: journalctl -u orama-namespace-olric@<name>.service --no-pager -n 50
  • Check WireGuard: wg show wg0 — look for recent handshakes and transfer bytes
  • Check gateway health: curl http://localhost:<port>/v1/health from the node itself
  • Node IPs: Check scripts/remote-nodes.conf for credentials, wg show wg0 for WG IPs