Commit 9b04ccc9 authored by intrigeri's avatar intrigeri
Browse files

Document another idea to use new hardware closer to its full potential

parent e4fb986d
......@@ -273,6 +273,25 @@ jobs. We might need to either make test jobs more robust on failure
([[!tails_ticket 17216]]),
or to start rebooting the Jenkins slaves after build jobs as well.
To utilize more fully such nodes, we may need to raise the number of
vCPUs allocated to TailsToaster. For example, assuming 8 hyperthreads
per box, and 2 VMs per box, with our current settings (2 vCPUs
allocated to TailsToaster and Cucumber eats another vCPU core):
- when running 1 single test suite job,
up to 3 hyperthreads out of 8 are used;
- when running 2 concurrent test suite jobs,
up to 2×3=6 hyperthreads out of 8 are used.
Raising this number may make the test suite run faster at least when
there's only 1 test suite job running. It might even make it faster,
in most scenarios, when 2 concurrent test suite jobs are running: they
generally won't be using that much CPU power at the same time. OTOH it
could introduce some test fragility.
We should first measure if it's worth the effort.
## Run builds and/or tests in the cloud
<div class="caution">
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment