[SimGrid-user] MSG_TRANSFER_FAILURE errors
c.enderlin at me.com
Tue Nov 5 22:31:41 CET 2013
I’m currently facing an issue with communications in MSG. Can you explain me the difference there is between MSG_TIMEOUT and MSG_TRANSFER_FAILURE ?
The documentation says: « Either the network is down or the remote host has been shutdown. » But it doesn’t seem to be the case in my example :
A node ‘A’ enters in a loop that sends 410 request tasks to 410 others nodes with MSG_task_isend(). Then, it enters in another loop that waits for those 410 matching answers ( with MSG_task_irecv() ).
During the send-loop, some receiver nodes start to send their answers to node ‘A’, but it will never receive them, because of MSG_TRANSFER_FAILURE errors that occurs for those answers.
I can’t understand why this happen … Do you have an idea ?
Thanks a lot in advance ! :-)
More information about the Simgrid-user