Skip to content
Pre-deploy rejected for missing /data volume despite attached volume; same release succeeds without pre-deploy
lekom
OP

a month ago

Staging deployments are blocked at PRE_DEPLOY_COMMAND before Railway creates the container. Builds succeed. The error says there is no volume attached at /data, although the service's existing volume is attached, READY, and mounted successfully by the running deployment.

Please investigate the platform-side volume validation for pre-deploy containers and advise on a permanent fix. Our current evidence suggests a runtime mount requirement may be applied to the separate pre-deploy container; this is a hypothesis, not a confirmed internal cause.

Affected resources

Project: teams-agent (97006eac-ceef-440c-862f-0dcaf41e1d6b)

Environment: staging (b9973f77-dd52-4a7a-a6ec-b8780f218425)

Service: ta-gateway-staging (d155786d-d1d4-4363-bd14-a94038a556cc)

Volume: 4c432e42-7b9b-4160-a6e5-ac15769e50f1, mounted at /data

Region: us-east4-eqdc4a; Dockerfile builder; runtime V2

Pre-deploy command: node /app/dist/predeploy.js

Exact deploymentEvents error

This deployment expects a volume mounted at /data, but the service has no volume attached. Railway did not create the container: it would have run without the volume, and data written there is lost on the next deploy. Attach a volume at /data, or remove the mount, and deploy again.

Timeline and comparison (October 6, 2026; UTC)

  • Last normal success: 96de8795-ac4d-4a6a-b83a-3d1039f89c83. Pre-deploy ran from 18:28:05.475 to 18:29:31.403 UTC.
  • First failure: fbb1964f-20d3-46cd-b9a9-a1f3a0460c5b. Pre-deploy failed between 18:31:19.925 and 18:31:20.373 UTC with the error above, before application logs.
  • The two deployments have identical build/deploy serviceManifest settings and volumeMounts: ["/data"]. requiredMountPath is null in both. The intervening source changes only updated agent instructions and skill metadata; no deployment configuration, Dockerfile, or pre-deploy code changed.
  • Retries, a fresh source deployment, and refreshing the existing volume attachment in place did not resolve it. We did not delete or replace the volume.

Same-release reproduction during recovery

We manually performed the exact release's migrations, executor image preparation, and graceful work drain, then temporarily cleared the separate pre-deploy command while retaining the runtime volume. Deployment ab4086cb-203e-49dc-9c2f-8fc88e854485 succeeded, with the existing /data volume mounted.

After restoring the normal pre-deploy command, the same source release (8a82cd09d3d725b2f3c777eaeaa36a3990443835) failed again in deployment f00de6b3-ebd8-4e70-b0e6-7f9a9c5fde80 at 18:48:57–18:49:02 UTC with the same error. A subsequent deployment, 7266ad99-7f77-4e6f-86f6-69643ead602c, also failed.

Impact and requested help

A recovered staging deployment is serving, but the normal deployment path remains blocked in our last verification. Production was not changed. We cannot safely leave pre-deploy disabled because it performs release preparation and drains active work.

Your documentation states that pre-deploy containers do not mount volumes: https://docs.railway.com/deployments/pre-deploy-command

Could you inspect these deployment events, confirm whether this is a pre-deploy volume-validation regression, and provide a fix that preserves the runtime volume and normal pre-deploy execution?

Solved

2 Replies

Status changed to Awaiting Railway Response Railway • about 1 month ago


chandrika
EMPLOYEE

a month ago

Hey! Thanks for the detailed report. We tracked this down to a recent change on our side and we're rolling it back now.

Your volume and its data are safe. Once the rollback is out, redeploying should go through with your volume mounted as usual.

If you detached your volume or deployed without it in the meantime, reply here before redeploying and we'll help make sure nothing written since then gets lost.


Status changed to Awaiting User Response Railway • about 1 month ago


chandrika
EMPLOYEE

a month ago

Hey! The rollback is out, and deploys with a volume and a pre-deploy command should go through again. Please redeploy and your volume will be mounted as usual.

Sorry for the disruption, and let us know if anything still fails.


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...