1274302 - please reimage talos-linux64-ix-009 as ubuntu 14.04

Status: RESOLVED FIXED

(Infrastructure & Operations :: RelOps: General, task)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

I have 5 loaner machine to run talos on taskcluster on our existing IX machines.

in order to run docker, we need a linux kernel of 3.10+, our 12.04 machines have 3.2.

I tried to upgrade with apt-get dist-upgrade, likewise do-release-upgrade- a lot got changed around (I am on kernel 3.5), but ubuntu is still 12.04 (despite reboots).  apt-get update always fails for me on some gpg errors, and it seems as though my hacking to get ubuntu updates probably messed up the apt servers.

I do have other boxes if it would be better to leave this one alone:
-009
-029
-049
-055
-069

once I get 1 box running what I need, I will look into imaging others as needed.

what I need is a fresh install of ubuntu 14.04 (I believe that is easily available) and a password for a user/root so I can install additional packages as needed.

Comment 1

[Amy Rich [:arr] [:arich]]

I put the vanilla 14.04 we have on the kickstart server on -009. Note that this installs puppet by default, but I've killed off the puppetize.sh process since it won't do the right thing here (we don't have puppet working for 14.04 talos machines).

Note for installing any future machines this way, when you kickstart these choose puppetagain 14.04 x64 and supply a BOGUS kickstart pw. It won't work if you supply none at all (it won't get past kickseed). Then log in with the kickstart root pw and kill of puppetize.sh and delete /etc/init.d/puppet. Root login is only permitted via console on a default install, so you also need to change the root pw, and then enable root ssh in /etc/ssh/sshd_config and restarted sshd.

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

thank you Amy, this machine is working great so far- if I have some green talos jobs overnight, I will ask for a few more to be reimaged in the morning with the goal of getting full coverage this weekend.

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

it if would be possible to get 2 more machines re-imaged, that would be really helpful!

Possibly:
-029
-049

or ideally all of them:
-009
-029
-049
-055
-069

for my experiment I need to run 192 hours of tests, so a few machines is critical to helping out with this.

Comment 4

[Amy Rich [:arr] [:arich]]

These have all been reimaged. It turns out I also needed to zero out /etc/rc.local, since that's where the puppetization script is inserted.