Uploaded image for project: 'RHEL'
  1. RHEL
  2. RHEL-15501

Services dying due a deadly signal are not generating core dumps anymore

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Normal Normal
    • rhel-9.5
    • rhel-9.2.0
    • systemd
    • None
    • systemd-252-33.el9
    • None
    • Moderate
    • sst_cs_plumbers
    • ssg_core_services
    • 26
    • 8
    • False
    • Hide

      None

      Show
      None
    • Yes
    • Red Hat Enterprise Linux
    • None
    • Enhancement
    • Hide
      .The default value for the `DefaultLimitCore` `systemd` configuration option is now set to `unlimited:unlimited`

      Previously, the default value for the `DefaultLimitCore` `systemd` configuration option was set to `0:infinity`. As a result, all processes started by `systemd` had a soft process limit for core files set to `0`, and no core files were created by default. However, the process adjusted the limit as required.

      With this update, the default value for `DefaultLimitCore` is set to `unlimited:unlimited`. As a result, the core file size is not limited by default. The default size of the crash dumps in the `/etc/systemd/coredump.conf` `systemd-coredump` component configuration file is `1GiB`. Note that you can gather crash dumps for sporadic crashes, but ensure that the use of disk space by crash dumps remains conservative.

      NOTE: The crash dumps stored by `systemd-coredump` are removed after 14 days if not used.
      Show
      .The default value for the `DefaultLimitCore` `systemd` configuration option is now set to `unlimited:unlimited` Previously, the default value for the `DefaultLimitCore` `systemd` configuration option was set to `0:infinity`. As a result, all processes started by `systemd` had a soft process limit for core files set to `0`, and no core files were created by default. However, the process adjusted the limit as required. With this update, the default value for `DefaultLimitCore` is set to `unlimited:unlimited`. As a result, the core file size is not limited by default. The default size of the crash dumps in the `/etc/systemd/coredump.conf` `systemd-coredump` component configuration file is `1GiB`. Note that you can gather crash dumps for sporadic crashes, but ensure that the use of disk space by crash dumps remains conservative. NOTE: The crash dumps stored by `systemd-coredump` are removed after 14 days if not used.
    • Done
    • None

      What were you trying to do that didn't work?

      Services dying due to some deadly signal being received or raised do not generate core dumps anymore, this is new to RHEL9.

      This is problematic because it completely prevents getting the root cause for such anomaly.

      Please provide the package NVR for which bug is seen:

      systemd-252

      How reproducible:

      Always

      Steps to reproduce

      1. Kill any system service, e.g. rsyslog, with sending a SEGV

      Expected results

      Coredump being generated

      Actual results

      No coredump

       

      Root cause

      This is due to having the following in /etc/systemd/system.conf:

      DefaultLimitCORE=0:infinity 

      This leads to having all services have a Soft CORE limit of 0, which disables dumping cores.

      This setting was commented out on RHEL8, causing "unlimited" Soft limit to be set.

              msekleta@redhat.com Michal Sekletar
              rhn-support-rmetrich Renaud Métrich
              systemd maint mailing list systemd maint mailing list
              Frantisek Sumsal Frantisek Sumsal
              Mugdha Soni Mugdha Soni
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Created:
                Updated:
                Resolved: