Long running dagger-engine experience | Dagger | Page 1

burnt loom Oct 8, 2025, 4:27 PM

#

We recently (1) upgraded from 0.16 to 0.18 and (2) stopped resetting the engine on dev-vms nightly

#

We have been seeing errors like "failed to find snapshot" or "failed to sync" show up randomly and unexpectedly

burnt loom Oct 8, 2025, 4:47 PM

#

For now, the Windows treatment (docker restart dagger-engine) seems to resolve these issues when they come up

#

looking into this more, the errors are since we upgraded (1.5months ago) while we stopped restarting (6months ago)

#

Is there a way for us to get more information about what's actually going wrong when this comes up?

#

Hopefully this is enough of a snippet to get some leads? This happens during the loading of directories from the host at the beginning of a build

): Directory! 11.0s
! failed to get content hash: failed to get snapshot: failed to sync: failed to get content hash xattr: no data available
╰─▼ upload . from r2lqi0myibo0ut91zjxsuq1uv

azure stream Oct 8, 2025, 5:33 PM

#

(drive-by guess) How does this intersect with the built-in cache pruning behavior? Perhaps there's a cache address that's aged out, but something still refers to it?

burnt loom Oct 8, 2025, 6:09 PM

#

yeah, we have some custom garbage collection config, but haven't changed that in a much longer time. Made another post about that, but didn't gain any insights from it https://discord.com/channels/707636530424053791/1422236640692404417

#Long running dagger-engine experience