Skip to content
PostgreSQL volume stuck on “Resizing volume” after resize to 5 GB
guinsuki
OP

a month ago

Hello Railway Support,

I upgraded my workspace to Hobby and resized my PostgreSQL volume from 500 MB to 5 GB using Live Resize.

The database is now online and the volume graph shows a 5 GB capacity, but the service still displays “Resizing volume…”.

Volume usage also increased from approximately 480 MB to 856 MB. PostgreSQL queries show:

  • All databases combined: approximately 146 MB.
  • WAL files: 352 MB.
  • Very few dead tuples, with autovacuum having run.

Could you verify whether the resize completed successfully, resolve the pending status, and help identify the remaining disk usage?

Previously, PostgreSQL crashed with “No space left on device”. Please preserve all existing data.

Project: charismatic-intuition

Service: Postgres

Volume: postgres-volume

Screenshot_1.png

Solved

3 Replies

Railway
BOT

a month ago

The resize completed: your Postgres volume now reads as 5,000 MB (5 GB, the Hobby maximum) with a ready state, and it is attached to the Postgres service, which is running on a successful deployment. Nothing is still pending on the volume itself, so the "Resizing volume…" label doesn't reflect a resize still in progress.

The jump in the usage graph right after the resize is expected. Resizing allocates filesystem metadata for the new capacity, which can add hundreds of MB to the reported figure, and usage can then creep up slowly afterwards even when your data hasn't changed. A redeploy of the service stops that background growth. The figure counts storage blocks allocated on the volume, not the live size of your data, so it will always settle above what Postgres and df report inside the container. Resizing does not touch the data stored on the volume.


Status changed to Awaiting User Response Railway • about 2 months ago


guinsuki
OP

a month ago

According to PostgreSQL’s own measurements, we have approximately:

  • 146 MB across all databases, including tables and indexes.
  • 352 MB in WAL files.
  • 498 MB total identified.

However, Railway’s Volume Usage reports approximately 856 MB.

Could you clarify what accounts for the remaining 358 MB? Is this entirely filesystem metadata and allocated blocks created by the resize? How can we verify that breakdown, and is this additional space included in billable storage usage?


Status changed to Awaiting Railway Response Railway • about 2 months ago


Railway
BOT

a month ago

The Volume Usage figure counts the storage blocks allocated on the volume at the storage layer, not the logical size of the files inside it, so it will always sit above what Postgres measures. The remaining gap comes from two things. First, the resize itself allocates filesystem metadata for the new capacity, which can add hundreds of MB right away and can keep growing slowly until the service is redeployed. Second, storage is allocated in 16 KB blocks while the filesystem inside the container frees space in 4 KB blocks, so partially freed blocks stay allocated. None of this is your data, and nothing is wrong inside the container.

You can't see this breakdown from inside the container, because the allocation happens beneath the container's filesystem. The closest check you can run is df against du on the data directory. If those two match each other and roughly match Postgres's figures, the gap is only at the storage layer. Running VACUUM FULL would not reclaim it.

Yes, it is billable. Volume storage is billed on this same usage figure, averaged over the billing period, at $0.15 per GB-month. So the roughly 358 MB gap comes to about $0.05 per month. The provisioned 5 GB size itself is not billed.


Status changed to Awaiting User Response Railway • about 2 months ago


Railway
BOT

a month ago

This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!

Status changed to Solved Railway • about 1 month ago


Welcome!

Sign in to your Railway account to join the conversation.

Loading...