Stalled Backups - Cannot Cancel

Support
1

Hi folks, I’m having an issue where I can’t complete a backup. I posted this question in Slack as well, but haven’t heard back yet:

I have KopiaUI running as a docker container using this compose backing up to a OneDrive repo. I got everything running and ran my first snapshot… it looks like all of the files sent, but the snapshot state has been stalled as “running” for the past half an hour. Is this normal?

As far as I can tell, the container isn’t actually doing anything since the cpu and network usage dropped to 0 when I closed the UI in my browser.

image (25)
image (25)1692×1179 255 KB

I see these lines are repeated in the backup logs:

10:41:17.596 checkpointed root k236c9cd35c3246de31755b4f652246d3
10:41:19.686 keeping 2024-01-19 09:11:19.5038031 -0500 EST retention: [incomplete] pins: []
10:41:19.686 keeping 2024-01-19 09:56:04.991297122 -0500 EST retention: [incomplete] pins: []
10:41:19.686 keeping 2024-01-19 10:41:17.596241903 -0500 EST retention: [incomplete] pins: []
11:25:52.184 checkpointed root k236c9cd35c3246de31755b4f652246d3
11:25:52.194 keeping 2024-01-19 09:11:19.5038031 -0500 EST retention: [incomplete] pins: []
11:25:52.194 keeping 2024-01-19 09:56:04.991297122 -0500 EST retention: [incomplete] pins: []
11:25:52.194 keeping 2024-01-19 10:41:17.596241903 -0500 EST retention: [incomplete] pins: []
11:25:52.194 keeping 2024-01-19 11:25:52.184326742 -0500 EST retention: [incomplete] pins: []

I also see that there are partial snapshots, and the ones with the error seem to be repeating. The errors appear to be related to unsupported entry type, even though I left the Ignore Unknown Directory Entries global policy defaulted to true.

[Screenshot would be here but I’m only allowed to include one image upload apparently]

  • If I try to cancel the snapshot, it just hangs in the same state indefinitely (I waited an hour.)
  • If I try to recreate the Kopia container, it still thinks its doing a snapshot and hangs there.
  • If I completely roll the container, clear out its data/cache/configs folder, and freshly configure the repo connection: when I restart it still thinks its in the middle of a snapshot.

The only way to recover is to roll the container, delete the content from the OneDrive, and try again… but using the same exact filter rules I now get different files triggering the unsupported entry type errors. Interestingly, the snapshot overview show the same number (11) of total errors, and the tooltip on the snapshot error icon shows the same number of errors (9) there as well:

[Screenshot would be here but I’m only allowed to include one image upload apparently]

Compose

version: "3.7"
services:
  kopia:
    image: kopia/kopia:latest
    hostname: wyse
    container_name: kopia
    restart: unless-stopped
    privileged: true
    cap_add:
      - SYS_ADMIN
    security_opt:
      - apparmor:unconfined
    devices:
      - /dev/fuse:/dev/fuse:rwm
    networks:
      - reverse-proxy-network
    command:
      - server
      - start
      - --disable-csrf-token-checks
      - --insecure
      - --address=0.0.0.0:51515
      - --server-username=coanghel
      - --server-password=<areallygoodpassword>
    volumes:
      # Mount local folders needed by kopia
      - /home/coanghel/kopia/config:/app/config
      - /home/coanghel/kopia/cache:/app/cache
      - /home/coanghel/kopia/logs:/app/logs
      - /home/coanghel/rclone/config:/app/rclone
      # Mount local folders to snapshot
      - /:/hostfs
      # Mount repository location
      # Mount path for browsing mounted snaphots
    environment:
        KOPIA_PASSWORD: <alsoareallygoodpassword>
        TZ: America/New_York
        USER: coanghel
networks:
  reverse-proxy-network:
    name: reverse-proxy-network
    external: true