teepod failed to start

[gldeng]

While trying to start ./teepod I encountered the following issue:

2025-01-13T02:30:01.295469Z  INFO load_config: Loading config file: ./teepod.toml
2025-01-13T02:30:01.318777Z  INFO supervisor_client: Failed to connect to supervisor at unix:./run/supervisor.sock, trying to start supervisor    
2025-01-13T02:30:01.372155Z  INFO supervisor_client: waiting for supervisor at unix:./run/supervisor.sock to start, attempt 1    
2025-01-13T02:30:01.574019Z  INFO supervisor_client: waiting for supervisor at unix:./run/supervisor.sock to start, attempt 2    
2025-01-13T02:30:01.776866Z  INFO supervisor_client: connected to supervisor at unix:./run/supervisor.sock    
2025-01-13T02:30:01.779650Z  INFO config{profile=release}: rocket::config: http2=true log_level="INFO" log_format=Compact cli_colors=auto workers=8 max_blocking=64 ident=Teepod ip_header="X-Real-IP" limits={"bytes": 8KiB, "data-form": 2MiB, "file": 1MiB, "form": 32KiB, "json": 1MiB, "msgpack": 1MiB, "string": 8KiB} temp_dir=/tmp keep_alive=10 shutdown.ctrlc=true shutdown.signals={"SIGTERM"} shutdown.grace=2 shutdown.mercy=3 shutdown.force=true
2025-01-13T02:30:01.779872Z  INFO routes{count=2}: rocket::route: name="prpc_post" rank=-1 method=POST uri=/api/<method> uri.base=/api uri.unmounted=/<method> location=teepod/src/main.rs:76
2025-01-13T02:30:01.780864Z  INFO routes{count=2}: rocket::route: name="prpc_get" rank=-1 method=GET uri=/api/<method> uri.base=/api uri.unmounted=/<method> location=teepod/src/main.rs:76
2025-01-13T02:30:01.781736Z  INFO fairings{count=1}: rocket::fairing: name="Shield" kind=liftoff, response, singleton
Error: Failed to run host API

Caused by:
    Failed to bind host API : IO error: Cannot assign requested address (os error 99)

[kvinwang]

check the port of host api in teepod.toml:

[host_api]
port = 10002

Choose another if it is already used.

An make sure the host system supports vsock.

[gldeng]

@kvinwang teepod.toml didn't include that configuration. I added it and it was still failing. Tried a few ports, all didn't work.

And I checked the os and got:

/opt/meta-dstack/build$ lsmod | grep vsock
vsock_loopback         12288  0
vmw_vsock_virtio_transport_common    57344  1 vsock_loopback
vmw_vsock_vmci_transport    49152  0
vsock                  61440  3 vmw_vsock_virtio_transport_common,vsock_loopback,vmw_vsock_vmci_transport
vmw_vmci              106496  1 vmw_vsock_vmci_transport

[gldeng]

@kvinwang seems adding the following configuration resolves the issue.

[host_api]
address="vsock:4294967295"

[kvinwang]

@kvinwang seems adding the following configuration resolves the issue.

[host_api]
address="vsock:4294967295"

This doesn't really work. VSOCK host support is not enabled on your system which is required.
The address must be "vsock:2"