MongoDB won’t start on CentOS 6: diagnosing the terminate() called error

On a fresh 64-bit CentOS 6 installation in October 2013, starting MongoDB with service mongod start aborted before the database came online. The useful lines from the startup output were:

terminate() called, printing stack (if implemented for platform):
... boost::filesystem::status ...
... mongo::ProcessInfo::checkNumaEnabled ...
... mongo::ProcessInfo::SystemInfo::collectSystemInfo ...
Got signal: 6 (Aborted).

What resolved it on this host

The original fix was a system package update with yum update. The working theory was a glibc or related library issue. The trace identifies where startup failed, and the update changed multiple packages, so the specific cause remains unconfirmed.

What to check if mongod aborts today

  1. Read the first error in the MongoDB log, before the stack trace. On a systemd host, also check sudo journalctl -u mongod -b --no-pager.
  2. Record the MongoDB version with mongod --version and the OS release with cat /etc/os-release on modern Linux. Check that the installed package is intended for that OS and architecture.
  3. If the logs point to permissions, a missing path, or a dependency, address that specific error. Back up the data and configuration before changing packages or moving to a supported host. Use MongoDB’s documented upgrade path for the installed version.

CentOS 6 is a legacy environment. The 2013 package update is historical context. For an active deployment, use MongoDB’s self-managed production notes to check platform support and version-specific Linux settings.


Discover more from Juzhax Technology

Subscribe to get the latest posts sent to your email.