ci(windows): runner C: drive is full (os error 112), Windows job fails before building #51

Closed
opened 2026-11-11 10:42:42 +00:00 by charles · 0 comments
Owner

Seen on commit a2ce504001 (PR #50, task/14-app-runtime), runs 75 and 76. The Linux job is green for the same commit and cargo xtask ci passes locally.

Both Windows jobs fail within seconds while compiling, before any test runs:

error: failed to write query cache to `\\?\C:\fcargo\fluxer-native-xtask\debug\incremental\xtask-...\query-cache.bin`: There is not enough space on the disk. (os error 112)
error: error writing dependencies to `C:\fcargo\fluxer-native\debug\deps\wgpu_core-9a00e83fc325c828.d`: There is not enough space on the disk. (os error 112)

Logs: release assets ci-log-75-windows and ci-log-76-windows.

The last green Windows job was run 74 (PR #49), about 20 minutes earlier, so the shared target dirs C:\fcargo\fluxer-native and C:\fcargo\fluxer-native-xtask (or something else on C:) filled the disk in between. Every Windows job on every branch will fail until space is freed on the runner.

To reproduce: re-run the Windows job for any branch.

Not fixed from the task branch: it needs space freed on the runner itself, and possibly a ci.yaml change (for example CARGO_INCREMENTAL=0 on Windows, or pruning the target dir when free space is low), which is a CI task of its own. PR #50 is blocked on this for its ready-to-merge label.

Seen on commit a2ce50400195495fd0f4cad514021fa7cdbea3c7 (PR #50, task/14-app-runtime), runs 75 and 76. The Linux job is green for the same commit and `cargo xtask ci` passes locally. Both Windows jobs fail within seconds while compiling, before any test runs: ``` error: failed to write query cache to `\\?\C:\fcargo\fluxer-native-xtask\debug\incremental\xtask-...\query-cache.bin`: There is not enough space on the disk. (os error 112) error: error writing dependencies to `C:\fcargo\fluxer-native\debug\deps\wgpu_core-9a00e83fc325c828.d`: There is not enough space on the disk. (os error 112) ``` Logs: release assets `ci-log-75-windows` and `ci-log-76-windows`. The last green Windows job was run 74 (PR #49), about 20 minutes earlier, so the shared target dirs `C:\fcargo\fluxer-native` and `C:\fcargo\fluxer-native-xtask` (or something else on C:) filled the disk in between. Every Windows job on every branch will fail until space is freed on the runner. To reproduce: re-run the Windows job for any branch. Not fixed from the task branch: it needs space freed on the runner itself, and possibly a ci.yaml change (for example `CARGO_INCREMENTAL=0` on Windows, or pruning the target dir when free space is low), which is a CI task of its own. PR #50 is blocked on this for its `ready-to-merge` label.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
charles/fluxer-native#51
No description provided.