sysbox-fs DeadlineExceeded error

[demidenkoalex]

Hi everyone!

I'm using Sysbox to run BuildKit for building container images. Previously, I used Kaniko. BuildKit builds use a ramdisk.
I've noticed that despite images being built in RAM, disk IOPS have increased. Why might this be happening?

Kernel version: 6.18
Sysbox version: 0.7.0

In absolute terms, throughput has barely changed, but the number of IOPS has grown severalfold, which suggests some kind of small disk operations are taking place. I can't figure out why.
As a result, some builds are now failing with the following error:

ERROR: Failed to cleanup volumes
ERROR: Job failed (system failure): Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: failed to pre-register with sysbox-fs: failed to pre-register container with sysbox-fs: rpc error: code = DeadlineExceeded desc = stream terminated by RST_STREAM with error code: CANCEL: unknown (exec.go:72:24s)

Could you please help me understand why this is happening?

[rogerhu]

Saw the same thing today and believe it's a deadlock issue -- see #998

If you install grpcurl (https://github.com/fullstorydev/grpcurl), you can validate whether the API returns:

sudo grpcurl -plaintext -authority dummy -d '{"Id": "test-container-1234","InitPid": 123456, "Ctime": "2026-07-14T14:42:00Z"}' unix:///run/sysbox/sysfs.sock protobuf.sysboxStateChannel.ContainerPreRegistration
{
  "Success": true
}

Also running kill -SIGQUIT <sysbox-fs PID>and peeking inside /var/log/sysbox-fs.log should confirm the same problem.