[Bug 1789875] [NEW] Failed to send coredump datagram: No buffer space available

[valentin]

Public bug reported:

lsb_release -rd:
Description:    Ubuntu 18.04.1 LTS
Release:        18.04

apt-cache policy systemd:
systemd:
  Installed: 237-3ubuntu10.3
  Candidate: 237-3ubuntu10.3
  Version table:
 *** 237-3ubuntu10.3 500
        500 http://no.archive.ubuntu.com/ubuntu bionic-updates/main amd64 Packages
        100 /var/lib/dpkg/status
     237-3ubuntu10 500
        500 http://no.archive.ubuntu.com/ubuntu bionic/main amd64 Packages


A program running inside a docker container crashes and is supposed to
generate a core dump. I am not sure if the docker part is relevant. The
program prints the message `Aborted (core dumped)` however `coredumpctl`
does sometimes not show the core dump.

In `/var/log/syslog` I see:
systemd[1]: Started Process Core Dump (PID 783/UID 0).
systemd-coredump[783]: Failed to send coredump datagram: No buffer space available
systemd-coredump[784]: Coredump file descriptor missing.

The core limit is correctly set to unlimited and I do get core dumps
when trying the same thing with a testing executable which just calls
`abort` in the same docker container. The program that fails to core
dump uses a lot of memory (5 to 10 GB) but there is enough disk space
(80 GB) and memory available.

Weird is that the core dump sometimes works and sometimes does not.
Maybe relevant is that the system is under heavy load over all.

** Affects: systemd (Ubuntu)
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1789875

Title:
  Failed to send coredump datagram: No buffer space available

Status in systemd package in Ubuntu:
  New

Bug description:
  lsb_release -rd:
  Description:    Ubuntu 18.04.1 LTS
  Release:        18.04

  apt-cache policy systemd:
  systemd:
    Installed: 237-3ubuntu10.3
    Candidate: 237-3ubuntu10.3
    Version table:
   *** 237-3ubuntu10.3 500
          500 http://no.archive.ubuntu.com/ubuntu bionic-updates/main amd64 Packages
          100 /var/lib/dpkg/status
       237-3ubuntu10 500
          500 http://no.archive.ubuntu.com/ubuntu bionic/main amd64 Packages


  A program running inside a docker container crashes and is supposed to
  generate a core dump. I am not sure if the docker part is relevant.
  The program prints the message `Aborted (core dumped)` however
  `coredumpctl` does sometimes not show the core dump.

  In `/var/log/syslog` I see:
  systemd[1]: Started Process Core Dump (PID 783/UID 0).
  systemd-coredump[783]: Failed to send coredump datagram: No buffer space available
  systemd-coredump[784]: Coredump file descriptor missing.

  The core limit is correctly set to unlimited and I do get core dumps
  when trying the same thing with a testing executable which just calls
  `abort` in the same docker container. The program that fails to core
  dump uses a lot of memory (5 to 10 GB) but there is enough disk space
  (80 GB) and memory available.

  Weird is that the core dump sometimes works and sometimes does not.
  Maybe relevant is that the system is under heavy load over all.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1789875/+subscriptions