Description of problem: As per https://bugzilla.redhat.com/show_bug.cgi?id=1834974#c4 running podman containers via systemd without PIDfile and Type=forking attributes isn't a supported configuration. Version-Release number of selected component (if applicable): # ceph --version ceph version 15.2.4-9.el8cp (fd4d62d568194c0dbb787e1845e65ed5c1de1b1f) octopus (stable) # rpm -qa cephadm cephadm-15.2.4-9.el8cp.x86_64 How reproducible: 100% Steps to Reproduce: 1. Deploy ceph with cephadm Actual results: # systemctl show ceph-b2692c62-c535-11ea-a48b-fa163e07380d.service|egrep '(Type|PIDFile)' Type=simple # systemctl cat ceph-b2692c62-c535-11ea-a48b-fa163e07380d.service # /etc/systemd/system/ceph-b2692c62-c535-11ea-a48b-fa163e07380d@.service # generated by cephadm [Unit] Description=Ceph %i for b2692c62-c535-11ea-a48b-fa163e07380d # According to: # http://www.freedesktop.org/wiki/Software/systemd/NetworkTarget # these can be removed once ceph-mon will dynamically change network # configuration. After=network-online.target local-fs.target time-sync.target Wants=network-online.target local-fs.target time-sync.target PartOf=ceph-b2692c62-c535-11ea-a48b-fa163e07380d.target Before=ceph-b2692c62-c535-11ea-a48b-fa163e07380d.target [Service] LimitNOFILE=1048576 LimitNPROC=1048576 EnvironmentFile=-/etc/environment ExecStartPre=-/bin/podman rm ceph-b2692c62-c535-11ea-a48b-fa163e07380d-%i ExecStart=/bin/bash /var/lib/ceph/b2692c62-c535-11ea-a48b-fa163e07380d/%i/unit.run ExecStop=-/bin/podman stop ceph-b2692c62-c535-11ea-a48b-fa163e07380d-%i ExecStopPost=-/bin/bash /var/lib/ceph/b2692c62-c535-11ea-a48b-fa163e07380d/%i/unit.poststop KillMode=none Restart=on-failure RestartSec=10s TimeoutStartSec=120 TimeoutStopSec=120 StartLimitInterval=30min StartLimitBurst=5 [Install] WantedBy=ceph-b2692c62-c535-11ea-a48b-fa163e07380d.target Expected results: # systemctl show ceph-b2692c62-c535-11ea-a48b-fa163e07380d.service|egrep '(Type|PIDFile)' PIDFile=/path/to/container/pidfile.pid Type=forking
Please specify the severity of this bug. Severity is defined here: https://bugzilla.redhat.com/page.cgi?id=fields.html#bug_severity.
Does this cause any issues beyond the overlays taking up extra disk space as they did in the ansible bug ( https://bugzilla.redhat.com/show_bug.cgi?id=1834974 )? I've tried testing to see if this happens in cephadm using a combination of repeated bootstrap and rm-cluster commands and it seems like the overlays are properly deleted when rm-cluster is run. Is this not the case for your setup? If you're also not seeing the overlay buildup, is there some other issue this causes we should be aware of?
The overlay growth issue may not be an issue for cephadm but the systemd configuration is still not what the podman team recommend/support
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Red Hat Ceph Storage 5.0 bug fix and enhancement), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2021:3294