Rdb load failed: Internal error when loading RDB file 5

[syphernl]

Describe the bug
This morning our Dragonfly instance crashed out of the blue and fails to load the RDB. As a result, it is stuck in a Crashloop.

To Reproduce
Steps to reproduce the behavior:

1. Start Dragonfly with an existing snapshot in /dragonfly/snapshots.
2. Wait for the server to initialize.
3. Observe the logs.
4. Error message appears:
   Rdb load failed: Internal error when loading RDB file 5.

Expected behavior
Dragonfly should load the snapshot without errors and continue serving requests (or skip the snapshot?)

Screenshots
N/A

Environment (please complete the following information):

• Kernel: Linux 6.8.0-83-generic x86_64
• Containerized?: Kubernetes (Dragonfly Operator v1.2.1)
• Dragonfly Version: 1.33.1

Reproducible Code Snippet
N/A

Additional context
Logs around failure:

I20250920 05:04:05.042048     1 snapshot_storage.cc:186] Load snapshot: Searching for snapshot in directory: "/dragonfly/snapshots"
I20250920 05:04:05.044055     1 server_family.cc:1320] Loading /dragonfly/snapshots/dump-2025-09-20T00:14:59-summary.dfs
E20250920 05:04:05.386683    11 server_family.cc:1371] Rdb load failed: Internal error when loading RDB file 5

Related
Could be related to #5273 (comment) where the disk was filled up because DF doesn't clear out old snapshots.

[romange]

@syphernl could you please share the snapshot?

[romange]

without thw snapshot we won't be able to help. please reopen if you provide additional information.

[syphernl]

Unfortunately I cannot provide this snapshot since it contains PII-data and I had to clear out the entire cache to get Dragonfly back up.

What I think happened:

• We have a cron setup every 15 minutes to make a snapshot
• The size of each snapshot is ~100MB
• At some point the PVC became full, the last backup could not be written fully
• Dragonfly was restarted (because the node restarted), attempted to load the last backup and failed because it was not complete.