New answers tagged hpc
I faced this problem, and the reason is I have multiple network interfaces (GE, IB) on the compute nodes, which are both reachable to the admin node. The admin node in the affected compute node is defined on a separate subnet as it is supposed to be with a different NIC.
It turns out that you have to have a cloud service created in Azure first, you can do this by doing the following: NEW -> COMPUTE -> CLOUD SERVICE
Rebooting the node ensures that the node is working properly before the next job is assigned to it. Also, when you run a node out of memory and start swapping it will slow down and may become unresponsive. In this case, they may be using something like IPMI to power cycle the node.
"Best" is dependent, of course, but there are a few well known ways to boost IO performance on AWS. RAID0-together a bunch of General Purpose SSD EBS volumes. A few, large, provisioned IOP EBS volumes. If you're doing sequential work, use the 640GB of Instance Local storage you have with that type. A combination of the above three points. All of the ...
Top 50 recent answers are included