How to fix /dev/tape: No such file or directory

[k3nd0x]

Hi @barcus First of all thanks for your work 🥇

But unfortanly I have an issue with the storage daemon on a virtual maschine. (Ubuntu 20.04 LTS) A backup job is failing after a few seconds there is an error "error wanted 2000 OK storage, got 2902 Bad storage"

The docker logs of the sd container are saying: chown: /dev/[n]st*: No such file or directory chown: /dev/tape: No such file or directory

It is happening with the newest alpine-psql and the newest ubuntu-psql images :(

[barcus]

Hi @k3nd0x, thanks for your feedback This message is just a warning, it's not related to your storage daemon issue, you can ignore it. (chown /dev/*) I will update docker-entrypoint to avoid this warning when device does not exist.

Could be great to know why you storage daemon exit. It looks like you use docker-compose to run the Bareos stack Can you run docker-compose up (without -d) ? in this way bareos-sd will display an error message. It should be link to bareos storage configuration or your docker-compose file if it exit instantly on start

[k3nd0x]

@barcus Thanks for your answer :) the SD is starting without any problems. status storage in bconsole says everything is fine. When i start my first backup job the SD container exits and stops working only the chown failes are showing up in the logs. The SD can read, write & execute to the mounted storage device.

additionally I have to say I installed the bareos-client directly on the maschine (without docker-container) The client is reachable and should not be the problem.

[barcus]

Can you please add this in your compose file (inside bareos-sd container) before launching an other backup job ?

command:
      - "/usr/sbin/bareos-sd"
      - "-u"
      - "bareos"
      - "-f"
      - "-d"
      - "200"

It will enable high level debug for storage daemon

[k3nd0x]

Before a backup job:

[barcus]

sorry, i would like to see the error when the container exits, during the backup job i imagine :)

[barcus]

because bareos-sd crashed (exit 139) as we can see in your first screenshot

[k3nd0x]

Hi, sorry for the late reply: This are the full logs after starting the container and after the first job:

[barcus]

Seems that your storage daemon does not speak to bareos-fd but bareos-dir. I don't know why but i think there is an issue in your configuration. Take a look on my own logs !

[k3nd0x]

Oh shi.. I found it. My client was unable to resolve the name of the bareos-sd ... The name of the SD is "bareos-sd" but my client is running directly on the maschine not in a docker container. So the DNS entry (or /etc/hosts entry) for bareos-sd was missing directly on the maschine. Sorry I bothered you with my config :) but thanks for your help. My backup jobs are running now 💯

[barcus]

fixed warning output for /dev/tape and /dev/[n]st* #153