Error MailboxError: transport failed with wandb

post by Dudubs on Aug 2, 2023

I am getting this error when trying to use wandb on a LambdaLabs instance:

MailboxError Traceback (most recent call last)
  in
    1 import wandb
----> 2 wandb.init(project=‘test’)

~/.local/lib/python3.8/site-packages/wandb/sdk/wandb_init.py in init(job_type, dir, config, project, entity, reinit, tags, group, name, notes, magic, config_exclude_keys, config_include_keys, anonymous, mode, allow_val_change, resume, force, tensorboard, sync_tensorboard, monitor_gym, save_code, id, settings)
   1164 if logger is not None:
   1165 logger.exception(str(e))
--> 1166 raise e
   1167 except KeyboardInterrupt as e:
   1168 assert logger

~/.local/lib/python3.8/site-packages/wandb/sdk/wandb_init.py in init(job_type, dir, config, project, entity, reinit, tags, group, name, notes, magic, config_exclude_keys, config_include_keys, anonymous, mode, allow_val_change, resume, force, tensorboard, sync_tensorboard, monitor_gym, save_code, id, settings)
   1145 except_exit = wi.settings._except_exit
   1146 try:
--> 1147 run = wi.init()
   1148 except_exit = wi.settings._except_exit
   1149 except (KeyboardInterrupt, Exception) as e:

~/.local/lib/python3.8/site-packages/wandb/sdk/wandb_init.py in init(self)
   790 run_start_handle = backend.interface.deliver_run_start(run._run_obj)
   791 # TODO: add progress to let user know we are doing something
--> 792 run_start_result = run_start_handle.wait(timeout=30)
   793 if run_start_result is None:
   794 run_start_handle.abandon()

~/.local/lib/python3.8/site-packages/wandb/sdk/lib/mailbox.py in wait(self, timeout, on_probe, on_progress, release, cancel)
   279 if self._keepalive and self._interface:
   280 if self._interface._transport_keepalive_failed():
--> 281 raise MailboxError(“transport failed”)
   282
   283 found, abandoned = self._slot._get_and_clear(timeout=wait_timeout)

MailboxError: transport failed

It is a fresh instance, I also tried to terminate and launch a new one, same problem.

It is working on other environments that I have (only on LambdaLabs it is not working).

I also tried to use another wandb account, same problem.