git clone https://github.com/intel/lkp-tests.git cd lkp-tests make install lkp help
# browse and select a job you want to run, for example, jobs/hackbench.yaml ls lkp-tests/jobs # install the common dependencies for lkp lkp install
lkp split-job lkp-tests/jobs/hackbench.yaml # output is: # jobs/hackbench.yaml => ./hackbench-1600%-process-pipe.yaml # jobs/hackbench.yaml => ./hackbench-1600%-process-socket.yaml # jobs/hackbench.yaml => ./hackbench-1600%-threads-pipe.yaml # jobs/hackbench.yaml => ./hackbench-1600%-threads-socket.yaml # jobs/hackbench.yaml => ./hackbench-50%-process-pipe.yaml # jobs/hackbench.yaml => ./hackbench-50%-process-socket.yaml # jobs/hackbench.yaml => ./hackbench-50%-threads-pipe.yaml # jobs/hackbench.yaml => ./hackbench-50%-threads-socket.yaml # install the remaining dependencies for the splited job lkp install ./hackbench-50%-threads-socket.yaml lkp run ./hackbench-50%-threads-socket.yaml
Specify disk partitions by defining hdd_partitions/sdd_partitions in host file named with local hostname and then lkp split-job will write the disk partitions information to split job file automatically.
Please note that disk partitions may be formatted/corrupted to run job.
echo "hdd_partitions: /dev/sda /dev/sdb" >> lkp-tests/hosts/$(hostname | sed -r 's/-[0-9]+$//g' | sed -r 's/-[0-9]+-/-/g') lkp split-job lkp-tests/jobs/blogbench.yaml # output is: # lkp-tests/jobs/blogbench.yaml => ./blogbench-1HDD-ext4.yaml # lkp-tests/jobs/blogbench.yaml => ./blogbench-1HDD-xfs.yaml # ... lkp install ./blogbench-1HDD-ext4.yaml lkp run ./blogbench-1HDD-ext4.yaml
To run your own benchmarks that are not part of lkp-tests, you can use mytest job.
lkp split-job lkp-tests/jobs/mytest.yaml # output is: # jobs/mytest.yaml => ./mytest-defaults.yaml lkp run ./mytest-defaults.yaml -- <command> <argument> ...
lkp result hackbench
Most test cases should install/run well in
There is however some initial support for:
As for now, lkp-tests still needs to run as root.
If you want to add support for your Linux distribution you will need an installer file which allows us to install dependencies per job. For examples look at: distro/installer/* files.
Since packages can have different names we provide an adaptation mapping for a base Ubuntu package (since development started with that) to your own distribution package name, for example adaptation files see: distro/adaptation/*. For now adaptation files must have the architecture dependent packages (ie, that ends with the postfix :i386) towards the end of the adaptation file.
You will also want to add a case for your distribution on sync_distro_sources() on the file lib/install.sh.
git checkout -b my-new-feature)
git commit -am 'Add some feature')
git push origin my-new-feature)