Skip to content

Startup and Upgrade Problems

Fix a MetaMiner Server that will not start, a port conflict, a MetaMiner Database that is not ready, or a database update that failed after an upgrade.

How startup works

When the metaminer Windows service starts, the MetaMiner Server (MMS) reads its configuration, checks the license, opens its port (default 49714), starts the MetaMiner Database (MMDB) connection and applies any pending database updates, then starts the pull engines. Until that finishes, clients see "The server ... is still starting up."

flowchart LR
  A[Service start] --> B[Read configuration]
  B --> C[Check license]
  C --> D[Open port 49714]
  D --> E[Connect MMDB and apply updates]
  E --> F[Start pulls]
  F --> G[Ready for clients]

The installers never start the services. After an install or upgrade, start MetaMiner from the MetaMiner Administrator (MMA). See Start and Stop the Server.

Image of the MMA main window after a failed start, with the start failure message and the server health indicator visible

../../images/help/mma-start-failed.png

The service will not start

The service stops right after it starts

Symptom. The MMA reports that MetaMiner failed to start, the server health indicator turns red, or the metaminer service shows Stopped in Windows Services shortly after you start it.

Cause. A step in the startup sequence failed. The most common: the license file is missing or expired, the port is in use, the MMDB is unreachable, or a database update failed.

Fix.

  1. Read the reason the MMA shows after the start attempt.
  2. Open server\logs\slc.log (lifecycle events) and server\logs\server.log, and look at the last lines before the stop. For problems before logging starts, read server\logs\server.start.log.
  3. Match the reason to an entry on this page.

What to send support. A log bundle created right after the failed attempt. Do not purge the logs first.

License not valid

Symptom. The service does not start and the logs report a license validation failure, or the MMA asks you to choose a license file.

Cause. The license file is missing, expired, or not valid for this installation. The MMS also re-checks the license daily and stops when it is no longer valid.

Fix. Apply a current license key. See Apply a License Key.

What to send support. The license file name and the license section of server.log.

Not enough memory to start

Symptom. When you click Start, the MMA warns "Your system's available memory is running low and is below the required amount for MMS."

Cause. The heap configured for the MMS is larger than the memory currently free on the server.

Fix. Free memory on the server, or lower Memory allocation for MetaMiner's JVM. See Performance and Memory.

What to send support. Not needed unless the warning appears with plenty of free memory.

Port in use

MetaMiner Server could not start: local TCP port is held by another process

Symptom. The MMA reports "MetaMiner Server could not start: local TCP port 49714 is held by PID ..." or "... is in use; the owning process could not be identified."

Cause. Another process owns the MMS port. It can be an old MetaMiner process that has not exited yet, or an unrelated program. The MMS retries the port for about 30 seconds before it gives up.

Fix.

  1. Find the owner:
    netstat -ano | findstr :49714
    Get-Process -Id <pid>
    
  2. If it is an old MetaMinerServer.exe, wait a minute or end that process, then start MetaMiner again.
  3. If it is another program, stop it or move MetaMiner to a free port in MMA Preferences > Settings > General > Server Port. Click Validate, save, and give every MMC user the new port. See Ports.

The installer reserves the configured MetaMiner ports so that Windows does not hand them to other programs automatically. The reservation does not stop a program that asks for the port explicitly.

What to send support. The MMA message, the netstat output, and slc.log.

MetaMiner Database not ready

The embedded MMDB does not start or is slow to start

Symptom. The MMS waits at startup, the logs report that the database is not ready, or the InfolytikPostgres service is stopped.

Cause. The embedded PostgreSQL service could not start (disk full, data folder damaged or locked by antivirus, port 49614 in use), or it is still recovering after an unclean shutdown. MetaMiner waits up to 4 minutes for the embedded database before it continues.

Fix.

  1. Check free space on the drive that holds server\data\mmdb\.
  2. Check the InfolytikPostgres service in Windows Services and start it if it is stopped.
  3. Read the PostgreSQL logs in server\data\mmdb\logs\ and server\logs\postgres.log.
  4. Exclude the server\data\mmdb\ folder from real-time antivirus scanning, if your policy allows.
  5. Start MetaMiner again from the MMA.

What to send support. The PostgreSQL logs, postgres.log, mmdb.log, and a log bundle.

An external PostgreSQL MMDB cannot be reached

Symptom. The logs report connection errors or authentication failures for the MMDB.

Cause. The external database is down, the host, port or credentials changed, or the network blocks the database port.

Fix. Test the connection from the MetaMiner server with the same account, correct the MMDB settings in the MMA, and start again. See External PostgreSQL.

What to send support. PostgreSQL version, the error text from mmdb.log or server.log, and a log bundle.

Database update failed after an upgrade

After an upgrade, the first start applies the database updates that the new release needs. Each update runs in its own transaction; a failed update is rolled back.

MMDB update failed and was rolled back

Symptom. server.log or mmdb.log contains "MMDB update #... failed and was rolled back. Later MMDB updates are held; the chain resumes here on the next start", or "MMDB is missing objects MetaMiner needs". The service may stay running, but some features fail; or the service stops during startup.

Cause. The database refused an update script. Typical reasons: the MMDB account does not own the schema or lacks rights on an external PostgreSQL, or the disk is full.

Fix.

  1. Do not edit the database by hand and do not restore an older release over the new one.
  2. Collect the information below and contact Infolytik Support.
  3. Apply the fix support gives you (for example a grant by your DBA, or freeing disk space).
  4. Restart the MetaMiner service. Held updates are retried only when the service starts again; fixing the cause alone does not apply them.
  5. After the restart, check server.log and mmdb.log for the same messages. No message means the updates completed.

What to send support.

  • A log bundle, including mmdb.log, server.log, slc.log and server.start.log from the first start after the upgrade.
  • The release you upgraded from and to.
  • The MMDB type (embedded or external) and, for external PostgreSQL, its version and the account MetaMiner uses.

Next steps