U d@sddlZddlZddlmZddlmZddlmZeZ dZ dZ dZ dZ deeeeeed ddZddZddZejdddZdS)N)closing) get_loggerzAddress already in usezSocket Timeoutz_tcp_store/num_membersz_tcp_store/last_memberXT) is_server server_addr server_port world_sizetimeoutwait_for_workersc CsP|dkr$|dkr$td|d||dkr>td|d|dkrJ|nd}|dkr\|}nt}td|d|d |d |d |d z 1, got server_port=z , world_size=z sever_port: z, specified, ignoring retrieszCreating c10d store on :z world_size : z is_server : z timeout(sec):  )seconds)Z host_nameportr Z is_masterr r zSuccessfully created c10d storezport: z already in use, attempt: [/]zon z, port: z already in use) ValueErrorloginfo get_free_portdistZTCPStoredatetime timedelta_check_full_rank RuntimeErrorstr_ADDRESS_IN_USEwarning) rr r r r r retriesattemptrstoreer$O/tmp/pip-unpacked-wheel-ua33x9lu/torch/distributed/elastic/utils/distributed.pycreate_c10d_storesN "      r&c Csx|td}||kr |tdz|tWnDtk rr}z&t|tkr`td|d|nW5d}~XYnXdS)Nrz ztimed out waiting for all z members to join) add_MEMBER_CHECKINset_LAST_MEMBER_CHECKINgetrr_SOCKET_TIMEOUT TimeoutError)r"r idxr#r$r$r%r    rc Cs2t}t||dW5QRSQRXdS)Nr)get_socket_with_portr getsockname)sockr$r$r%rks r)returnc Cstjddtjtjd}|D]x}|\}}}}}t|||}z|d|d|WStk r}z|tj d|dW5d}~XYqXqt ddS) a Returns a free port on localhost that is "reserved" by binding a temporary socket on it. Close the socket before passing the port to the entity that requires it. Usage example :: sock = _get_socket_with_port() with closing(sock): port = sock.getsockname()[1] sock.close() # there is still a race-condition that some other process # may grab this port before func() runs func(port) localhostN)hostrfamilytype)r3rrzSocket creation attempt failed.)exc_infozFailed to create a socket) socket getaddrinfo AF_UNSPEC SOCK_STREAMbindlistenOSErrorcloserrr)addrsaddrr5r6proto_sr#r$r$r%r/qs"   "r/)rrrTr)rr8 contextlibrZtorch.distributedZ distributedrZ'torch.distributed.elastic.utils.loggingrrrr,r(r*boolrintfloatr&rrr/r$r$r$r%s2    B