New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 732520 link

Starred by 0 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug

Blocked on:
issue 733394



Sign in to add a comment

amd64-generic simple chrome builder required on PFQ

Project Member Reported by achuith@chromium.org, Jun 12 2017

Issue description

We have a FYI builder in the chromeos.chrome waterfall:
https://uberchromegw.corp.google.com/i/chromeos.chrome/builders/chrome-tot-chromeos-amd64-generic

We need an identical builder in the PFQ.

There's a proposal to make the PFQ the source of LKGM. We use the LKGM version in the chromium tree for the simple chrome builders, and we ought to vet this version in the PFQ.

This is an important pre-condition to adding VM tests to the simple chrome builders.

As an example, the current version 9588.0.0 which hasn't updated in ~2 weeks has a broken VM.
 
Cc: akes...@chromium.org
Components: Infra>Client>ChromeOS
Labels: -Pri-1 Pri-2
We do currently build Simple Chrome in the PFQ:

https://uberchromegw.corp.google.com/i/chromeos/builders/amd64-generic-chromium-pfq/builds/10206/steps/TestSimpleChromeWorkflow/logs/stdio

We could add the VMTest stage to that builder as well.

Because amd64-generic doesn't run hardware tests, it probably would not affect the total run time of the PFQ.

That said, it might be 'simpler' to have a separate builder for this, at least from a gardening perspective. Configuring a new builder is probably a lot less simple.

The VMTest stage would require this to be a bare-metal builder (I'm not sure what it is right now). This approach would give us a high degree of confidence that the VM would be functional, and that simple chrome would build, so I'm fine with this if it's the simplest thing to do. 

I'll note that this still wouldn't be identical to the chromium simple chrome builder which would run a functionally similar but different test (vm_sanity.py vs the smoke suite), and would be a GCE instance instead of a baremetal builder. There's also a real possibility of breakage due to the differences in timing from running non-hw accelerated VMs in the GCE instance.
Labels: Hotlist-LKGM
Owner: achuith@chromium.org
Status: Assigned (was: Untriaged)
Adding build-configs to cbuildbot is a self-serve softy. ;)
Treating this as a task-bug filed by achuith for work he's (already?) doing.
If you want infra help (beyond reviews / consult), re-add to infra triage by un-assigning.


I'm not sure I understand the requirements completely here. So, we do TestSimpleChromeWorkflow on the PFQ. So what's left? Is the request here to add VMTest stage to this (and other PFQ) builders?

fwiw, amd64-generic-chromium-pfq runs on a baremetal build-slave: https://cs.corp.google.com/chromeos_public/chromite/cbuildbot/waterfall_layout_dump.txt?l=4
Blockedon: 733394
Components: Infra>Client>ChromeOS>CI
Components: -Infra>Client>ChromeOS

Sign in to add a comment