Volume grew 20GB → 80GB overnight

[waverlycabinets]

I’m running a Remix app on Fly.io that grew from 20GB to 80GB overnight. We’re running litestream and passing backups to Tigris Object Storage. Does anyone know what might be happening? It started two days ago.

[mayailurus]

Hi… The first thing I’d try is to SSH in and see which parts of the filesystem tree are taking up the most space.

$ fly m start  # ensure at least one Machine running
$ fly ssh console
# apt-get update
# apt-get install --no-install-recommends ncdu
# df -h /data  # replace with the volume's mount point
# ncdu  /data

(Assumes a Debian-based image.)

That last command will let you interactively drill down into directories (via a TUI) to see which sub-directories, sub-sub-directories, etc., are unexpectedly bulky.

For example, sometimes people have the volume also store user uploads, which would tend to stand out from this perspective.

Also, it would be prudent to check the Avail reported by df, since the numbers that you gave are very round (exact multiples of 10)—and might be the volume’s capacity instead of its actual usage. The discrepancy between those two can complicate the detective work, particularly when the amount of real data was large in the past but now is small again…

https://fly.io/docs/reference/configuration/#auto-extend-volume-size-configuration

Hope this helps a little!