Re: archive rebuilds
On Sun, May 12, 2013 at 4:01 AM, Lucas Nussbaum <lucas@debian.org> wrote:
> Hi!
Hello Lucas,
> First, thanks all for volunteering!
You are very welcome! I look forward to helping out with this task.
> From now on, let's use the debian-qa@ list for coordination (so, please
> subscribe, if not done already)
I am already subscribed to the debian-qa@ list.
> I'll try to walk you through the tools used to do rebuilds and bug
> filing. There are three big steps.
I would like to thank you for putting so much time into documenting
and training us for this task. I looked over the links you provided in
an earlier email, and I was very impressed with how thorough the
documentation for all of the scripts/tasks was.
> Step 1: Use AWS to run all the tests and generate logs
> Step 2: Scan the logs (that's done automatically), generate list of
> failures. Also upload the logs to some public place
> Step 3: Look at each failure and submit bugs (that's done
> semi-automatically)
>
> In that mail, I will address Step 1. Once at least two of you have achieved it,
> I'll move to Step 2.
I'll give Step 1 a try sometime today (once my SSH key gets added).
> There's some documentation (not completely up-to-date) on
> http://wiki.debian.org/qa.debian.org/ArchiveTesting
> For Step 1, most of the instructions are in
> http://anonscm.debian.org/gitweb/?p=collab-qa/cloud-scripts.git;a=blob_plain;f=README;hb=HEAD
>
>
> In the context of this "tutorial", your 'mission' is to rebuild all
> packages that take between 100 and 110 seconds to build, and to get the
> build logs for them.
>
> First, I need your SSH key, so that you can connect to AWS.
Please find my public SSH key attached to this message.
> Then, connect to our "master" node on EC2. That node must never be shut
> down. I recommend you add sthing like:
> Host ec2
> User root
> Hostname ec2-50-17-69-153.compute-1.amazonaws.com
> to your .ssh/config.
>
> In order to make your life slightly easier (and to reduce a bit the risks in
> case of mistakes :) ), I've already reserved 5 "medium" instances for you. So
> you don't need to do the resources reservation steps of README. The list of
> those nodes is in /root/cloud-scripts/nodes.m1.medium.tutorial.
> It also means that you must run masternode with the --no-halt switch, so that
> nodes are not shut down are the end of the rebuild.
>
> Then, everything happens in /root/cloud-scripts/
>
> Read README, generate a list of tasks, check that you can connect to the
> "builder" nodes, and rebuild those packages. In two of you happen to follow
> that tutorial at the same time, generate a list of tasks with an ID so that log
> names don't conflict.
>
> Please let me know if you have questions (either by mail or on IRC, #debian-qa).
>
> When you are done, take a look around (esp. at masternode and process-tasks),
> and let me know, so that we can move to Step 2.
>
> Lucas
Thanks,
Nathan
£É6ü?
Êû;
uÈw˜bec2.pubQ?šÃssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQC9u/4gEyiNPmn1/9hx2Zd+ACZcE8WgZU3uyR6Vc7P0HPYb7tZ8CCkkP36k2PshwdPaT+XVa15S4+1muJ7VJzGawDDcTol35HK7sAefyCu3dYwwYlKTiXJCpfpkIOspY3/gQi/e6h40582pRlk/Hi9s905aJxOnd7AIMJCaRNiECjVdxTxighbWd3fvvRrpexx3Cyt/4rKyUE5nFZAB8wOQRpLl8kWLWb6MdFptPmDdwI4ubNj5DIg2+F4rm8zKBlbG4LxpJNC3X7ZhuXHCXri6uf1KJ5KzSdkWhy6UkELwBJ39akMPHzz/tJRjLreBPXZ7BALBKTvsjIXBAngi/GVL nhandler@kirk
‰ Q?šÃ