Details
-
Task
-
Resolution: Unresolved
-
Minor
-
None
-
None
-
9223372036854775807
Description
This README file is out of date: https://git.whamcloud.com/?p=fs/lustre-release.git;a=blob;f=lustre/tests/README;h=00634a77eb61ae98021baab7938cf797d46d7b67;hb=HEAD
Now the process of running tests is as follows:
I normally do what you are tying to do, which is: reserve nodes on trevis or onyx, use loadjenkinsbuild to load the Lustre kernel/RPMs onto the nodes, set up Lustre, modify the Lustre test configuration file to match the Lustre setup and then launch tests.
When you get an allocation on trevis/onyx, you are getting an allocation of VMs and I run on those VMs. As the README states, you can use tests/llmount.sh to easily create a single node Lustre file system using loopback devices. Sometimes this is all you need to answer some questions. Usually you will need several nodes (vms) and setup Lustre in a different configuration.
Of course, this is about how we run tests inside the company.
Need to check the README file and make them relevant for external users.