My pipeline has mysteriously broke but I haven't changed it | Dagger | Page 1

dapper schooner Dec 20, 2024, 11:07 PM

#

For a while it mysteriously failed before even starting, until finally I changed a random line and it was able to advance. Almost like the module itself was cached and the cached version was broken.

Now it fails at the start of the pipeline, always:

│ ✘ .stdout: String! 17.8s
│ ! failed to compute cache key: failed to read expected number of bytes: unexpected EOF
│ │ ✔ magicache layer download 0.2s
│ │ ✔ magicache layer download 0.2s
│ │ ✔ magicache layer download 0.2s
│ │ ✔ magicache layer download 0.1s
│ │ ● magicache layer download 0.4s

https://dagger.cloud/cordialsys/traces/0209b4e23d46265227e65fb756e5f623

I'm pretty stumped as I haven't changed the pipeline at all and has been pretty stable until today.

I've tried resetting docker, restarting, and disabling Rosetta virtualization to no avail

dagger v0.15.1 (registry.dagger.io/engine:v0.15.1) darwin/arm64

Is there a way I can get it to stop caching?

daring mortar Dec 21, 2024, 1:15 AM

#

@dapper schooner are you running this in an OSS repo by any chance?

dapper schooner Dec 21, 2024, 1:19 AM

#

no unfortately, but i export something to share

daring mortar Dec 21, 2024, 1:20 AM

#

no need

#

I'll reset your cloud cache now and I think that'll solve the issue

#

I believe this is coming froma very tricky bug we've been hunting for a while

#

gimme 1 sec

#

@dapper schooner can you try again please?

dapper schooner Dec 21, 2024, 1:26 AM

#

seems to be working now

daring mortar Dec 21, 2024, 1:26 AM

#

🙏

#

please ping me if it happens gain

dapper schooner Dec 21, 2024, 1:28 AM

#

okay will do, I can share a zip next time if it'd be helpful

#My pipeline has mysteriously broke but I haven't changed it