/opt/netbox/docker-entrypoint.sh: line 23: wait: pid is not a child of this shell

[tkloda]

Hello,

I'm having an issue spinning up NetBox instance in OpenShift. I'm using RHEL7 based image instead of default Alpine Linux. Everything seems to be fine except to the point where NetBox tries to connect to the database from OpenShift pod.

I can see this error in the logs: /opt/netbox/docker-entrypoint.sh: line 23: wait: pid is not a child of this shell
The piece of code in docker-entrypoint.sh file:

while [ "${CUR_DB_WAIT_TIME}" -lt "${MAX_DB_WAIT_TIME}" ]; do
  # Read and truncate connection error tracebacks to last line by default
  exec {psfd}< <(./manage.py showmigrations 2>&1)
  read -rd '' DB_ERR <&$psfd || :
  exec {psfd}<&-
  wait $! && break
  if [ -n "$DB_WAIT_DEBUG" ]; then
    echo "$DB_ERR"
  else
    readarray -tn 0 DB_ERR_LINES <<<"$DB_ERR"
    echo "${DB_ERR_LINES[@]: -1}"
    echo "[ Use DB_WAIT_DEBUG=1 in netbox.env to print full traceback for errors here ]"
  fi
  echo "⏳ Waiting on DB... (${CUR_DB_WAIT_TIME}s / ${MAX_DB_WAIT_TIME}s)"
  sleep "${DB_WAIT_TIMEOUT}"
  CUR_DB_WAIT_TIME=$((CUR_DB_WAIT_TIME + DB_WAIT_TIMEOUT))
done
if [ "${CUR_DB_WAIT_TIME}" -ge "${MAX_DB_WAIT_TIME}" ]; then
  echo "❌ Waited ${MAX_DB_WAIT_TIME}s or more for the DB to become ready."
  exit 1
fi

The problematic line is line 23 wait $! && break . It seems, because of the above pid issue, the line will never break out of the loop and the program will time out.

The log below is repeating till it times out:

⏳ Waiting on DB... (0s / 30s)
--
  | /opt/netbox/docker-entrypoint.sh: line 23: wait: pid 29 is not a child of this shell
  | ߧ젬oaded config '/etc/netbox/config/configuration.py'
  | ߧ젬oaded config '/etc/netbox/config/extra.py'
  | ߧ젬oaded config '/etc/netbox/config/logging.py'
  | ߧ젬oaded config '/etc/netbox/config/plugins.py'
  | admin
  | [ ] 0001_initial
  | [ ] 0002_logentry_remove_auto_add
  | [ ] 0003_logentry_add_action_flag_choices
  | auth
  | [ ] 0001_initial
  | [ ] 0002_alter_permission_name_max_length
  | [ ] 0003_alter_user_email_max_length
  | [ ] 0004_alter_user_username_opts
  | [ ] 0005_alter_user_last_login_null
  | [ ] 0006_require_contenttypes_0002
  | [ ] 0007_alter_validators_add_error_messages
  | [ ] 0008_alter_user_username_max_length
  | [ ] 0009_alter_user_last_name_max_length
  | [ ] 0010_alter_group_name_max_length
  | [ ] 0011_update_proxy_permissions
  | [ ] 0012_alter_user_first_name_max_length
  | circuits
  | [ ] 0001_squashed (1 squashed migrations)
  | [ ] 0002_squashed_0029 (28 squashed migrations)
  | contenttypes
  | [ ] 0001_initial
  | [ ] 0002_remove_content_type_name
  | dcim
  | [ ] 0001_squashed (1 squashed migrations)
  | [ ] 0002_squashed (1 squashed migrations)
  | [ ] 0003_squashed_0130 (129 squashed migrations)
  | [ ] 0131_consoleport_speed
  | [ ] 0132_cable_length
  | [ ] 0133_port_colors
  | extras
  | [ ] 0001_squashed (1 squashed migrations)
  | [ ] 0002_squashed_0059 (58 squashed migrations)
  | [ ] 0060_customlink_button_class
  | [ ] 0061_extras_change_logging
  | [ ] 0062_clear_secrets_changelog
  | ipam
  | [ ] 0001_squashed (1 squashed migrations)
  | [ ] 0002_squashed_0046 (45 squashed migrations)
  | [ ] 0047_prefix_depth_children
  | [ ] 0048_prefix_populate_depth_children
  | [ ] 0049_prefix_mark_utilized
  | [ ] 0050_iprange
  | sessions
  | [ ] 0001_initial
  | taggit
  | [ ] 0001_initial
  | [ ] 0002_auto_20150616_2121
  | [ ] 0003_taggeditem_add_unique_index
  | tenancy
  | [ ] 0001_squashed_0012 (12 squashed migrations)
  | [ ] 0002_tenant_ordering
  | users
  | [ ] 0001_squashed_0011 (11 squashed migrations)
  | virtualization
  | [ ] 0001_squashed_0022 (22 squashed migrations)
  | [ ] 0023_virtualmachine_natural_ordering
  | ⏳ Waiting on DB... (3s / 30s)
  | /opt/netbox/docker-entrypoint.sh: line 23: wait: pid 33 is not a child of this shell

Does anyone already encountered such issue? What would be a proper way to fix it?

[tobiasge]

I moved this into a discussion, because it is not a bug with our image.

I assume the bash in RHEL 7 is older than that in Alpine and doesn't understand the syntax that is used. Why do you need a RHEL7 based image? We're also running our Netbox in Openshift with the Alpine based image.

[tkloda]

Hey @tobiasge
unfortunately I'm forced to use RHEL based image, because our platform does not support images from other Linux distributions. I guess some modification must be made to the docker-entrypoint.sh so that RHEL7 understands it. I still don't how but I guess there must be a way how to overcome this :)

[tkloda]

I was able to fix it by modifying line 23 in docker-entrypoint.sh like this:
exec {psfd}<&- & wait $! && break

Bash version: 4.2.46(2)-release

Now I encountered another issue, but that should be easy to fix as I don't have unit installed yet:

✅ Initialisation is done.
--
  | /opt/netbox/launch-netbox.sh: line 48: exec: unitd: not found
  | ⏳ Waiting for control socket to be created... (1/10)

@tobiasge Thank you for pointing me to the right direction!