Start-up and access

The XPLG service does not start or keeps restarting, the console does not open, HTTPS and certificate errors, the wrong Java version, out-of-memory restarts, Linux file limits, X8 roles that stay unhealthy.

What you will do: get from "it does not come up" to a console you can sign in to.

The service does not start, or starts and stops

CheckWhereWhat to do
The process loglog/xpologlog and the Servlet Container logs under the configuration directory; journalctl -u XPLG (Linux service); the Windows Event LogThe first error after start usually names the cause: a port in use, a missing directory, a corrupt lax line
Ports 30303 / 30443 already takenss -ltnp (Linux), netstat -ano (Windows)Stop the other process or change the ports in ServletContainer/conf/server.xml
Configuration directory not writableSettings > General path; file permissionsThe XPLG user needs full read and write on the external configuration directory and the data paths
A broken lax editXpoLog.sh.lax / XpoLog.lax, line LAX.NL.JAVA.OPTION.ADDITIONALEvery flag on that one line, separated by spaces; a stray quote or newline stops the JVM
Linux limits too lowulimit -n, ulimit -u as the XPLG user64,000 open files and 20,000 processes in /etc/security/limits.conf; log out and in; restart
Service runs the instance manually during a patchrunXpoLog.sh statAfter a patch, stop the manual process and start the service again

X8: docker compose ps shows a role that is not healthy; docker compose logs <role> names the cause; verify.sh checks the whole stack. First boot can take one to eight minutes. /node/data on NFS, cgroups v1, a missing memlock 256 MB or an .env with wrong SELF_ALIAS / seed IPs are the usual reasons.

Out-of-memory restarts

  • The heap is too small for the daily volume, or so large that the host swaps. Set -Xmx to about 60 % of RAM on Linux and 70 % on Windows, at most 32 GB per process (System requirements).
  • Keep -XX:+ExitOnOutOfMemoryError -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=… so the process restarts and leaves a dump for support.
  • System Health > Memory shows the trend; a steady climb points at a parsing or monitor configuration rather than at size.

The console does not open

  1. The service is running (above) and listening: curl -k https://<host>:30443/logeye/ returns HTML.
  2. A firewall or security group allows 30303 / 30443 from your network.
  3. Behind a load balancer: sticky sessions are on, htmlUtil.BaseUrl and mail.link.baseUrl point at the public name, health checks hit /logeye/.
  4. HTTPS only: the HTTP port was set to 0 on purpose; use https://.

Certificate and HTTPS errors

  • The default certificate is self-signed: the browser warns until you install a CA-signed .jks or .pfx in ServletContainer/conf/server.xml with the full chain, sslProtocol="TLSv1.2" and, for .pfx, certificateKeystoreType="PKCS12".
  • After replacing the keystore, restart. A wrong password or alias leaves the HTTPS port closed; the Servlet Container log says which.

Wrong Java version

System Health > System Information lists the Java version per node. Below 17 means the Java 17 migration has not been applied (patch 9001); current patches refuse to deploy until it is. The NoJava installer variant needs LAX_VM pointing at your own Java 17.

Cannot sign in

  • Security was activated and the default account disabled: use the administrator you created; a lost password is reset from the internal realm by another administrator.
  • SSO: the identity provider's response must match the SAML / OIDC settings under Settings > Security; check the audit log for the rejected assertion.
  • Idle timeout is 30 minutes by default.

If none of this applies, export the diagnostics (Diagnostics) and write to support@xplg.com.

Next: Data collection

Source: XPLG wiki, pages 1594131986, 1594132180, 1594132942, 1594132938, 3150544899, 3151659011, 1708130305, 3161718811 (https://xpolog.atlassian.net/wiki/spaces/XPOL/pages/1594132938).