Infrastructure Workgroup?

Infrastructure Workgroup?

[Andrew Geissler]

Based on some recent discussion in this area, I'm contemplating
forming an Infrastructure workgroup. The focus would be on improving
our CI infrastructure and the scripts and tools around it.

There's been a lot of great community effort already around our CI
build scripts.

Some thoughts on goals:
- Documenting the existing flow
- Alternatives to existing flow (ability to use freely provides clouds
for open source projects, ...)
- Managing the Jenkins nodes - Bringing new nodes online quickly
(chef, puppet, ansible?)
- Managing and optimizing the CI flow (repository CI, meta CI, openbmc CI)
-- Which systems do we build?
- Process for adding hardware CI (mechanism for individuals to add HW
CI to gerrit commits)
- Optimizations to existing scripts (docker building, CI, QEMU)
- Gerrit, Jenkins, and build node periodic update process
- Overall maintenance and responsibilities for infrastructure

Thoughts or interest?

Andrew

Re: Infrastructure Workgroup?

[Adriana Kobylak]

On 2018-10-30 10:49, Andrew Geissler wrote:
> Based on some recent discussion in this area, I'm contemplating
> forming an Infrastructure workgroup. The focus would be on improving
> our CI infrastructure and the scripts and tools around it.
> 
> There's been a lot of great community effort already around our CI
> build scripts.
> 
> Some thoughts on goals:
> - Documenting the existing flow
> - Alternatives to existing flow (ability to use freely provides clouds
> for open source projects, ...)
> - Managing the Jenkins nodes - Bringing new nodes online quickly
> (chef, puppet, ansible?)
> - Managing and optimizing the CI flow (repository CI, meta CI, openbmc 
> CI)
> -- Which systems do we build?
> - Process for adding hardware CI (mechanism for individuals to add HW
> CI to gerrit commits)
> - Optimizations to existing scripts (docker building, CI, QEMU)
> - Gerrit, Jenkins, and build node periodic update process
> - Overall maintenance and responsibilities for infrastructure
> 
> Thoughts or interest?

+1 . This would become important as the org grows. Also setting 
guidelines
on the minimum tests and platforms that need to pass before a change is
accepted.

> 
> Andrew

Re: Infrastructure Workgroup?

[Sai Dasari]

On 10/30/18, 8:55 AM, "openbmc on behalf of Andrew Geissler" <openbmc-bounces+sdasari=fb.com@lists.ozlabs.org on behalf of geissonator@gmail.com> wrote:

    Based on some recent discussion in this area, I'm contemplating
    forming an Infrastructure workgroup. The focus would be on improving
    our CI infrastructure and the scripts and tools around it.
    
    There's been a lot of great community effort already around our CI
    build scripts.
    
    Some thoughts on goals:
    - Documenting the existing flow
    - Alternatives to existing flow (ability to use freely provides clouds
    for open source projects, ...)
    - Managing the Jenkins nodes - Bringing new nodes online quickly
    (chef, puppet, ansible?)
    - Managing and optimizing the CI flow (repository CI, meta CI, openbmc CI)
    -- Which systems do we build?
    - Process for adding hardware CI (mechanism for individuals to add HW
    CI to gerrit commits)
    - Optimizations to existing scripts (docker building, CI, QEMU)
    - Gerrit, Jenkins, and build node periodic update process
    - Overall maintenance and responsibilities for infrastructure
    
    Thoughts or interest?
Thanks for starting this great initiative! Will be happy to be part of this infra workgroup.    
    Andrew
    

Re: Infrastructure Workgroup?

[krtaylor]

On 10/30/18 10:49 AM, Andrew Geissler wrote:
> Based on some recent discussion in this area, I'm contemplating
> forming an Infrastructure workgroup. The focus would be on improving
> our CI infrastructure and the scripts and tools around it.

Actually, when I was suggesting this, I was thinking even broader. There 
is a need for folks interested in keeping *all* of our services and 
system resources healthy, not just our test systems.

I would think that the Test WG would take care of defining and 
automating test, the infrastructure WG would take care of the resources 
that Test needed for CI.

That said, it makes sense for the first order of work for the Infra WG 
would be to bolster the CI Test services and resources. There will be 
big early overlaps for sure.

I'll be happy to help as I can, at least initially. I have CI/DevOps 
experts in my network that I can get recommendations from as well.

Kurt Taylor (krtaylor)

> 
> There's been a lot of great community effort already around our CI
> build scripts.
> 
> Some thoughts on goals:
> - Documenting the existing flow
> - Alternatives to existing flow (ability to use freely provides clouds
> for open source projects, ...)
> - Managing the Jenkins nodes - Bringing new nodes online quickly
> (chef, puppet, ansible?)
> - Managing and optimizing the CI flow (repository CI, meta CI, openbmc CI)
> -- Which systems do we build?
> - Process for adding hardware CI (mechanism for individuals to add HW
> CI to gerrit commits)
> - Optimizations to existing scripts (docker building, CI, QEMU)
> - Gerrit, Jenkins, and build node periodic update process
> - Overall maintenance and responsibilities for infrastructure
> 
> Thoughts or interest?
> 
> Andrew
> 

Re: Infrastructure Workgroup?

[Joel Stanley]

On Wed, 31 Oct 2018 at 02:24, Andrew Geissler <geissonator@gmail.com> wrote:
>
> Based on some recent discussion in this area, I'm contemplating
> forming an Infrastructure workgroup. The focus would be on improving
> our CI infrastructure and the scripts and tools around it.
>
> There's been a lot of great community effort already around our CI
> build scripts.
>
> Some thoughts on goals:
> - Documenting the existing flow
> - Alternatives to existing flow (ability to use freely provides clouds
> for open source projects, ...)
> - Managing the Jenkins nodes - Bringing new nodes online quickly
> (chef, puppet, ansible?)
> - Managing and optimizing the CI flow (repository CI, meta CI, openbmc CI)
> -- Which systems do we build?
> - Process for adding hardware CI (mechanism for individuals to add HW
> CI to gerrit commits)
> - Optimizations to existing scripts (docker building, CI, QEMU)
> - Gerrit, Jenkins, and build node periodic update process
> - Overall maintenance and responsibilities for infrastructure
>
> Thoughts or interest?

I'm interested.

I wrote the scripts that were used as the basis for our CI. I recently
revisited them to fix an issue, and there is a lot of bitrot there. I
would be interested in seeing that resolved.

I would also like to help enable the compute power we do have to be
used effectively. Currently we have a lot of CPU cores sitting idle
while there are jobs in the queue.

Cheers,

Joel

Re: Infrastructure Workgroup?

[Brad Bishop]

> On Oct 30, 2018, at 11:49 AM, Andrew Geissler <geissonator@gmail.com> wrote:
> 
> Thoughts or interest?

Thanks for doing this Andrew.  I’d like to participate.

-brad

Re: Infrastructure Workgroup?

[Ping Mao]

Hi,  Brad:

Where can I check your replies?

Thanks,
Ping

On 11/2/18, 2:46 AM, "openbmc on behalf of Brad Bishop" <openbmc-bounces+pmao=linkedin.com@lists.ozlabs.org on behalf of bradleyb@fuzziesquirrel.com> wrote:

    
    > On Oct 30, 2018, at 11:49 AM, Andrew Geissler <geissonator@gmail.com> wrote:
    > 
    > Thoughts or interest?
    
    Thanks for doing this Andrew.  I’d like to participate.
    
    -brad