Host lookup faliure

[skathaitrooney]

Hello Experts,

Head node: prodhn01
Compute node: prodcn01

My job is meant to run on the above head node and the compute node.

Sometimes i am getting this error:

APT_PMaddrInfoFor(): getaddrinfo(host = prodcn01, port = ) failed with error -3 - Temporary failure in name resolution: node prodcn01 cannot be used

When i re run the job after a minute or so, job runs fine on the same compute node.

According to my understanding i think host lookup faliure must be the issue that the DNS was not able to resolve. I went to the DNS team, but they couldnt find any errors/warnings on their system.

Also, sometimes DNS lookup fails even for the head node, i did not get why would the i get the same error for the head node even if my job is running on the same head node:

APT_PMaddrInfoFor(): getaddrinfo(host = prodhn01, port = ) failed with error -3 - Temporary failure in name resolution: node prodhn01 cannot be used

Any ideas??

I am thinking to update the hosts file to overcome this issue...

[ray.wurlod]

Since you have not specified what name resolution technique you are using it's hard to comment but, yes, hard coding IP address to name mapping in the hosts file should successfully resolve this issue.