MongoDB 8.2.5 (tcmalloc-google) SIGSEGV crash loop on Linux kernel 6.19 — stable on 6.18.7

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Duplicate
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • None
    • ALL
    • Hide

       1. Run MongoDB 8.2.5 (mongo:8 or mongo:8.2.5) via Docker on a host with Linux kernel 6.19.x
        2. Start the container with a persistent data volume
        3. Observe container exits with code 139 (SIGSEGV) within seconds of startup
        4. Container restarts (Docker restart policy) and crashes again in ~31 seconds
        5. Check coredumps: `coredumpctl list | grep mongod`
        6. Check Docker events: `docker events --filter container=<name> --filter event=die`

        To confirm kernel regression:
        7. Downgrade host kernel to 6.18.7
        8. Start the same container with the same data volume — MongoDB starts cleanly

      Show
       1. Run MongoDB 8.2.5 (mongo:8 or mongo:8.2.5) via Docker on a host with Linux kernel 6.19.x   2. Start the container with a persistent data volume   3. Observe container exits with code 139 (SIGSEGV) within seconds of startup   4. Container restarts (Docker restart policy) and crashes again in ~31 seconds   5. Check coredumps: `coredumpctl list | grep mongod`   6. Check Docker events: `docker events --filter container=<name> --filter event=die`   To confirm kernel regression:   7. Downgrade host kernel to 6.18.7   8. Start the same container with the same data volume — MongoDB starts cleanly
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      MongoDB 8.2.5 enters a persistent SIGSEGV crash loop (exit code 139, signal 11) when running
        on Linux kernel 6.19. The process restarts every ~31 seconds via Docker restart policy, and
        each unclean crash triggers a full WiredTiger journal recovery on the next startup.

        The crash loop ran undetected for ~10 days, producing 1,705+ coredumps and 2,000+ WiredTiger
        journal recovery cycles. On a btrfs filesystem, this caused ~60GB of CoW extent accumulation
        (ghost storage growth of ~10GB/hr).

        After downgrading to kernel 6.18.7-200.fc43.x86_64, MongoDB started cleanly and ran
        indefinitely with zero crashes on identical hardware, config, and data.

        Environment:
        - MongoDB: 8.2.5 (tcmalloc-google allocator)
        - Docker: running via docker compose
        - OS: Fedora 43
        - Kernel (broken): 6.19.x
        - Kernel (stable): 6.18.7-200.fc43.x86_64
        - Filesystem: btrfs
        - Hardware: bare metal homelab server

        Observations:
        - `docker events` showed `die exit=139` every ~31 seconds
        - `coredumpctl list` captured 1,705+ MongoDB coredumps
        - THP set to `always` had no effect — crash loop continued unchanged
        - Zero crashes on kernel 6.18.7 with identical MongoDB config and data volume
        - MongoDB startup log on 6.18.7 showed clean starts with no journal recovery messages

            Assignee:
            Unassigned
            Reporter:
            Ali Alrabeei
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: