I hereby claim:
- I am cgwalters on github.
- I am walters (https://keybase.io/walters) on keybase.
- I have a public key whose fingerprint is 1CEC 7A9D F7DA 85AB EF84 3DC0 A866 D7CC AE08 7291
To claim this, I am signing this object:
I hereby claim:
To claim this, I am signing this object:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
This is a GPG (clear)signed document that I will keep updated with my public SSH keys for general-purpose use.
My GPG key can be found on https://keybase.io/walters (fingerprint: 1CEC7A9DF7DA85ABEF843DC0A866D7CCAE087291)
If you want to give me access to a machine, I recommend using the raw
view, and
gpg --verify
on the document to verify my signature (and
not just github's copy):
Mostly more fighting with OpenStack instances. Played around with the new QEOS7, hit random issues. Tried a bit more to bootstrap OpenShift containerized on OS1 Public, but the lack of DNS was problematic. Realized I had bootstrapping issues trying to build/update a dnsmasq container.
Ended up backing off a multi-node cluster and just set everything to point at one VM for playing around. I/O appears to be really, really slow.
However, I did successfully use the Let's Encrypt Docker image, which worked out quite nicely.
Update: Apparently d-s-s failed because my rootfs was too small, left me with a corrupted pool, had to blow away all of my images after adding a cinder volume.
I'm trying to set up my own OpenShift cluster under aos.verbum.org, running on CentOS Atomic Host, inside an OpenStack cluster.
The containerized install is giving me all of the pain of docker-vs-systemd that we've known about forever.
Spent about an hour trying to track down why my attempts to add Cinder PVs were being rejected. It turns out that there's currently a hard requirement that the hostname match the instance ID in the cloud provider...which is rather annoying, as I have my hostname set to the public version.
The goal here is to create a local mirror of the upstream OSTree repository.
Log into a shell on the host, and run the tools container. This isn't
strictly necessary right now as we could use the host's ostree
binary
too, but in the future the tools container might contain more scripts
for things like this. Once we're inside the tools container, we make
an unprivileged user, as general best practice. It also shows that
OSTree is perfectly happy to mirror/manipulate content as non-root.
From 252613fff3142dc2be7e9a3c47fc9c4488ff75da Mon Sep 17 00:00:00 2001 | |
From: Colin Walters <[email protected]> | |
Date: Thu, 31 Mar 2016 21:27:24 -0400 | |
Subject: [PATCH 1/2] Demote pytest/coverage to Fedora only | |
We aren't running the tests on CentOS due to dependency issues (whole | |
vast swath of Python modules to backport), so add these deps into %if | |
fedora as well. | |
--- | |
python-docker-py.spec | 2 +- |
We want to replace Colin's Homu with one more collaboratively maintained in a public OpenShift v3 instance, but for now:
openssl rand -hex 20
for secret-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
This is an assertion that I am a maintainer of the https://github.com/ostreedev/ostree project. Its most recent release: https://github.com/ostreedev/ostree/releases/tag/v2016.5 is signed by cgwalters, aka https://keybase.io/walters
I assert that I am also the user "walters" on freenode.
On Friday I did: https://ci.centos.org/job/atomic-fedora-ws/
See: https://pagure.io/fork/walters/workstation-ostree-config/commits/f23
Notes on my personal conversion:
Hit grub2.cfg not being updated when doing install-inside-existing
Need to detect this case and update both grub.cfg
?
Need atomic run centos/tools
for strace on flatpak
In Project Atomic we implemented two models to install what one might call "system containers".
There's now "atomic install" and "atomic install --system". I will call the former "atomic install" and the latter "system containers".
These both use: