Mounting and performance

Mount

mkdir -p ~/mnt/teldrive
rclone mount teldrive: ~/mnt/teldrive

For applications that need seek-heavy writes, use VFS caching:

rclone mount teldrive: ~/mnt/teldrive \
  --vfs-cache-mode full \
  --vfs-cache-max-size 20Gi \
  --vfs-cache-max-age 24h

Size the cache to available local disk.

Chunk size

Default:

chunk_size = 512Mi

Accepted range: 64 MiB–2000 MiB. Larger chunks reduce API calls but use more memory per concurrent upload.

Example:

[teldrive]
type = teldrive
api_host = https://drive.example.com
api_key = YOUR_API_KEY
chunk_size = 256Mi
upload_concurrency = 2

Concurrency

These controls multiply:

• rclone --transfers;
• backend upload_concurrency;
• Teldrive telegram.upload-threads.

Start conservatively:

rclone copy /srv/data teldrive:/archive \
  --transfers 2 \
  --checkers 4 \
  --progress

Raise one setting at a time while monitoring memory, CPU, Telegram errors, and PostgreSQL load.

Hashing

hash_enabled = true

Keep BLAKE3 hashing enabled unless you have a measured reason to disable it.

Upload host

Optional dedicated upload endpoint:

upload_host = https://uploads.example.com

It must expose the same compatible Teldrive API/authentication.

Teldrive encryption

encrypt_files = true

Requires server-side content-encryption keys. See Content encryption.

If rclone lsf teldrive: fails, fix authentication/connectivity before debugging mount settings.