Openjdk-infrastructure: Need a 2nd create_installer_linux node for redundency

[andrew-m-leonard]

build-scaleway-ubuntu1604-x64-1 is the only node for create_installer_linux
if that has an issue, eg.Agent goes offline, disk issue, or whatever, then the builds hang...

Please can we have a 2nd one for redundency?

[sxa]

@Haroon-Khel As requested, a new Ubuntu16.04 is now available if you want to progress this: Your key can access [email protected]

[Haroon-Khel]

I have run the playbook on the machine. Ive given it the name build-godaddy-ubuntu1604-x64-1, as its original hostname was just its ip address. Am running an installer job on the machine as a test.

[sxa]

That's a bit odd - it shouldn't have just been the IP address unless you'd run the playbook with the IP in the hosts file previously

[Haroon-Khel]

build-godaddy-ubuntu1604-x64-1 is now in jenkins. For the sake of testing its compatibility with the create_installer_linux job, I created a temporary gpg key on the machine. The existing installer machine, build-scaleway-ubuntu1604-x64-1, has an adopt key on the machine. Is this a unique key that needs to be on all installer machines to sign rpm and deb packages? Or should I just create a new key for build-godaddy-ubuntu1604-x64-1? @sxa

Standalone create installer linux job which has passed on build-godaddy-ubuntu1604-x64-1
https://ci.adoptopenjdk.net/job/build-scripts/job/release/job/standalone_create_installer_linux/121/console

[aahlenst]

It needs to be the same key everywhere. And it should not change.

[Haroon-Khel]

The new key is now on the machine. Ive added the fpm and ubuntu labels, so the machine should get used by the create_installer_linux job. Lets see if it survives the nightlies

[sxa]

Doesn't look like that's the right set of tags as the job is still only trying to run on https://ci.adoptopenjdk.net/computer/build-scaleway-ubuntu1604-x64-1/:

[sxa]

Closing this since it's part of https://github.com/AdoptOpenJDK/openjdk-build/issues/1044