Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-32407

Systemd processes not being moved to cpuset/systemd.slice

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Undefined Undefined
    • None
    • 4.14
    • Node Tuning Operator
    • +
    • Important
    • No
    • CNF Compute Sprint 252, CNF Compute Sprint 253
    • 2
    • False
    • Hide

      None

      Show
      None
    • Hide
      *Cause*: cpuset-configure.sh script might run before all system processes are created.
      *Consequence*: system services are not being moved to the custom system slice
      *Fix*: Triggering the script to run only once crio is initialized. since crio is initialized at the very end of the boot order just before kubelet, this will guarantee no late starting system processes will fall between the cracks.
      *Result*: Bug doesn’t present anymore.
      Show
      *Cause*: cpuset-configure.sh script might run before all system processes are created. *Consequence*: system services are not being moved to the custom system slice *Fix*: Triggering the script to run only once crio is initialized. since crio is initialized at the very end of the boot order just before kubelet, this will guarantee no late starting system processes will fall between the cracks. *Result*: Bug doesn’t present anymore.
    • Bug Fix
    • In Progress

      This is a clone of issue OCPBUGS-31692. The following is the description of the original issue:
      —
      This is a clone of issue OCPBUGS-30569. The following is the description of the original issue:
      —
      Description of problem:

      This is only applicable to systems that install a performance profile

There seems to be a race condition where all systemd spawed processes are not being moved to /sys/fs/cgroup/cpuset/system.slice.

This is suppose to be done by the one-shot cpuset-configure.service. 
Here is a list of processes I see on one lab that are still in the root directory 

/usr/bin/dbus-broker-launch --scope system --audit
dbus-broker --log 4 --controller 9 --machine-id 071fd738af0146859d2c04b7fea6d276 --max-bytes 536870912 --max-fds 4096 --max-matches 131072 --audit
/usr/sbin/NetworkManager --no-daemon
/usr/sbin/dnsmasq -k
/sbin/agetty -o -p -- \u --noclear - linux
sshd: core@pts/0

      Version-Release number of selected component (if applicable):

          4.14, 4.15

      How reproducible:

      Steps to Reproduce:

          1. Reboot a SNO with a peformance profile applied 
    2.
    3.

      Actual results:

      Expected results:

      Additional info:

              rh-ee-rbaturov Ronny Baturov
              openshift-crt-jira-prow OpenShift Prow Bot
              Mallapadi Niranjan Mallapadi Niranjan
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated:
                Resolved: