When the service’s failed within the installation process, one cannot view what really caused the crashes. The only way to see what happend is to dig out the Host system journal. I found a handmade command line to do so:
journalctl -N
returning all available fields
_SOURCE_MONOTONIC_TIMESTAMP _STREAM_ID CODE_FUNC PRIORITY SYSLOG_TIMESTAMP _EXE _SYSTEMD_SLICE _SOURCE_REALTIME_TIMESTAMP _SYSTEMD_CGROUP NM_DEVICE NM_LOG_DOMAINS CONTAINER_ID_FULL SYSLOG_FACILITY _SYSTEMD_INVOCATION_ID CODE_FILE CONTAINER_NAME MESSAGE CODE_LINE _CAP_EFFECTIVE _CMDLINE CONTAINER_TAG _TRANSPORT _COMM TIMESTAMP_BOOTTIME NM_LOG_LEVEL TIMESTAMP_MONOTONIC _PID _MACHINE_ID SYSLOG_PID _HOSTNAME _GID SYSLOG_IDENTIFIER _UID _SYSTEMD_UNIT CONTAINER_ID _BOOT_ID INVOCATION_ID UNIT_RESULT COMMAND UNIT EXIT_CODE EXIT_STATUS MESSAGE_ID LIMIT_PRETTY MAX_USE_PRETTY _UDEV_SYSNAME DISK_AVAILABLE MAX_USE JOURNAL_PATH _KERNEL_SUBSYSTEM CURRENT_USE AVAILABLE_PRETTY DISK_KEEP_FREE SYSLOG_RAW JOURNAL_NAME AVAILABLE CURRENT_USE_PRETTY _KERNEL_DEVICE LIMIT DISK_AVAILABLE_PRETTY DISK_KEEP_FREE_PRETTY _UDEV_DEVNODE SEAT_ID
To display the journal of my container before it’s started, I’ve got to invoke :
FIELD=CONTAINER_NAME; journalctl -F $FIELD | xargs printf "journalctl $FIELD=%s; " | bash
Interesting information outputs the following:
May 23 .. f955ecddc48f[816]: [info] Applying target state
May 23 ..f955ecddc48f[816]: [warn] Ignoring unsupported or unknown compose fields: envFile
May 23.. f955ecddc48f[816]: [warn] Ignoring unsupported or unknown compose fields: envFile
May 23 .. 955ecddc48f[816]: [event] Event: Service start {"service": {"appId":1564784,"serviceId":414920,"serviceName":"MYSERVICENAME","r>
May 23 ..f955ecddc48f[816]: [info] Internet Connectivity: OK
Do you know there a better way to get more information about how the services boot up ?