Hi all,
Just to quickly let you know that
console-openshift-console.apps.ocp.ci.centos.org was switched to point
to new oauth2 settings.
So please try to login with "CentOS_and_Fedora_account", that is
pointing to new id.centos.org, itself now using (Free)IPA backend that
is is the common/merged one between Fedora and CentOS.
Reminder : if your account wasn't existing in FAS and so just imported
from ACO, you have first to reset your password through new portal
https://accounts.centos.org (or https://accounts.fedoraproject.org, as
it's the same backend, same settings, just different visual them).
Verify your group membership and/or projects ACLs in openshift, and if
you have troubles, create ticket on
https://pagure.io/centos-infra/issues so that Vipul and David can have a
look and eventually correct things (if you have a new email address or
nickname from your FAS/ACO merged account)
Cheers,
--
Fabian Arrotin
The CentOS Project | https://www.centos.org
gpg key: 17F3B7A1 | twitter: @arrfab
Hi All !
As announced multiple times (including but not limited to
https://lists.centos.org/pipermail/centos-devel/2021-February/076442.html),
Fedora and CentOS will merge authentication soon.
It was already merged for Staging environment, where SIGs contributors
could test things and now it's time to really merge
https://accounts.centos.org and https://admin.fedoraproject.org/accounts
on the new system.
Let me point you first to the mail sent to Fedora so please read it
first to have a little bit of background/history :
https://lists.fedoraproject.org/archives/list/announce@lists.fedoraproject.…
As you can see, the Fedora migration will happen next week.
Based on current timeline and agenda, we'll proceed like this for the
CentOS migration :
* Friday April 2nd :
* We'll "freeze" https://accounts.centos.org in Read-only mode
* Fedora infra team launches the fas2ipa script to import centos
users/groups not existing (yet) in new IPA setup (if you had a fedora
account matching your account in accounts.centos.org, you'll not be
imported again, but rather be added to your imported centos groups - so
merged -)
* Monday April 5th
* quick sanity check for the import script result and some internal
checks, then
* Real CentOS infra authentication switch : it's hard to give a
timeline but we'll start with https://cbs.centos.org (I'll announce
downtime in separate mail when we'll have full agenda) and then proceed
with the other services.
How will you be impacted ?
If you use any kind of service authenticated by either TLS cert from
https://accounts.centos.org (that's the case for cbs.centos.org, or mqtt
notifications), you'll *have* to retrieve a new cert. (more information
in the SIGGuide will appear on due time.
Same for services using authentication tied to
https://accounts.centos.org through https://id.centos.org (for
openid/openidc, etc)
So this mail doesn't contain all the information for how to retrieve new
TLS cert, how to reset password, etc but more to give you the date when
we'll have smallest possible downtime while reconfiguring system to
switch to new authentication (FWIW, all changes were automated through
ansible for our staging environment, so we'll just reapply same process
for the production one)
Have a nice week-end !
--
Fabian Arrotin (all excited to finally see this project arriving at
deploy time) :)
The CentOS Project | https://www.centos.org
gpg key: 17F3B7A1 | twitter: @arrfab
Hi All,
Last year we announced a new OCP4 cluster deployment with a request to
migrate workloads to it (from openshift 3.6). I have been working with
a few projects to help migrate and have received good feedback.
On March 1st 1200 UTC, we will shut down the legacy cluster. If you
have a job running there, please open a ticket on our tracker[0] so
that we can work with you to migrate them on time.
[0] https://pagure.io/centos-infra/new_issue?template=ci-migration
Thank You
--
Vipul Siddharth
He/His/Him
Fedora | CentOS CI Infrastructure Team
Hi All,
There is a security update coming today in jenkins. I will update the
ci.centos.org jenkins plugins (and other instances that I maintain) as
the severity is High (check [0])
I will disable jobs execution at 2pm UTC today and update+restart at
2:30-3 pm UTC
All jobs requested will be queued in that period.
If you have a jenkins instance, please consider doing the same.
Thank You
[0] https://groups.google.com/d/topic/jenkinsci-advisories/sEeqkNwz5IQ
--
Vipul Siddharth
He/His/Him
Fedora | CentOS CI Infrastructure Team
Hi all,
We seem to have issue with an IBM Power 8 used within CI and so we have
to re-balance CI nodes that you can request through Duffy API for
ppc64/ppc64le.
My question sounds more like a survey : I think that most (if not all)
CI projects actually still building (and testing in CI) just target the
ppc64le architecture (Little Endian) and so not the ppc64 (Big Endian) one.
We'd like to hear from you and depending on the needs, we can
eventually drop ppc64 architecture for CI tests, and so have more
(re-balanced) ppc64le resources .
Opinions ?
--
Fabian Arrotin
The CentOS Project | https://www.centos.org
gpg key: 17F3B7A1 | twitter: @arrfab
Hi All,
Just wanted to inform you that today at 1330 UTC, I will be updating
ci.centos.org jenkins to version 2.263.2 to catch up with new
CVE/vulnerabilities fixes.
If you are using a namespace in Openshift cluster, please consider
updating your jenkins version as well.
--
Vipul Siddharth
He/His/Him
Fedora | CentOS CI Infrastructure Team
Sending it here too, but ideally if you have questions/comments, that
would be better to join the thread on centos-devel list.
Thanks !
-------- Forwarded Message --------
Subject: [CentOS-devel] CentOS/Fedora authentication system merge
(Please Read)
Date: Wed, 27 Jan 2021 08:58:13 +0100
From: Fabian Arrotin <arrfab(a)centos.org>
Reply-To: The CentOS developers mailing list. <centos-devel(a)centos.org>
To: The CentOS developers mailing list. <centos-devel(a)centos.org>
# Introduction and background
As it was preannounced some time ago , the CentOS Board agreed to merge
the CentOS accounts (https://accounts.centos.org) with the Fedora FAS
(https://admin.fedoraproject.org/accounts/)
As both projects were running their own instance of FAS (running on
el6/CentOS 6, so coming to EOL, so that needed to be migrated to new
solution/platform), but that there are a lot of contributors common to
both projects, it made sense to "migrate and merge" both into one, and
so having only one account that can be used for both.
The AAA/Noggin team worked in the last months on the new authentication
system that will be used as foundation.
The core block will be (Free)IPA (https://www.freeipa.org , already
available in the distribution) and the community portal feature will be
provided by noggin (https://github.com/fedora-infra/noggin)
If you want to know more about noggin, consider watching the
presentation given at last Fedora Nest event
(https://www.youtube.com/watch?v=x1SevUmkE60)
# What does it mean for you, contributors and SIG members ?
Fedora already had an IPA infra, but "hidden" behind FAS, so accounts
were already created in IPA backend.
For CentOS, we were just using plain FAS, so users in our own backend
(fas db).
The "Merge" operation will go like this :
- Fedora will kick fas2ipa script
(https://github.com/fedora-infra/fas2ipa), synchronizing FAS attributes
back into IPA, including group memberships coming from FAS/Fedora
- Then the same process will be ran but importing from ACO
(https://accounts.centos.org) into the same IPA backend.
That's where the "fun" begins:
* If the same nick/account exists at both side, the script is
considering FAS as authoritative (remember, the FAS user *already*
exists there, and is only modified for group[s] membership and attributes)
* What is used to consider same nick/account being the same person ?
the email (validated when registering account) will be used as primary
key. So that means that you should *now* verify/update your email
address in FAS and ACO so that they match
* in case of a email address mismatch, the ACO account isn't migrated
(group membership) but put in a queue to be verified
* in case of matching email address, existing account is added to
imported ACO groups
The "open" question is about what to do for same account but in fact
being different people (question is debated between Fedora and CentOS
through the AAA initiative)
# What has been already done ?
You can follow publicly the status through dedicated tracker (
https://github.com/orgs/fedora-infra/projects/6 ), but let me focus on
the CentOS Side (sending this to centos-devel so centos contributors)
In the last months, Fedora already deployed a staging (.stg.) IPA
instance, as well as a noggin community portal.
For CentOS, we deployed (to be able to test integration) the following
components in front of the Fedora IPA:
* https://accounts.stg.centos.org (using noggin, with a centos visual
theme applied)
* https://id.stg.centos.org (ipsilon, used for openid/openidc IdP)
We then reached out to some "key users" to validate that some
applications migrated to new authentication system were working fine.
We tested with :
* pagure (https://git.stg.centos.org)
* koji
* openshift/OCP
* some other apps using openid
In December 2020, there was a first ran of the fas2ipa script, so
(consider this a snapshot) existing accounts in both FAS and ACO were
merged.
>From that import, there were 123 accounts that were duplicates ones, but
as said, it can be that they are the same account but using different
email addresses.
# What do you have to do ?
You can try to login through https://accounts.stg.centos.org and see if
you can login.
Important remark: if you *didn't* have a FAS account , your account was
imported/created for the first time in IPA, so that means that you'll
have to use the "Forgot Password ?" feature on portal to reset your
account (mail will be sent to email address tied to your account)
# When will the real migration happen ?
We'll wait on AAA/noggin team to give us estimated date, and when
they'll migrate Fedora first.
Once that will be done, we'll migrate ACO to the new setup (probably
fas2ipa script ran during a week-end, but to be announced)
# How will that impact my workflow for CentOS as SIG member ?
Worth knowing that all deployed services using ACO will have to be
reconfigured for AAA.
That currently means :
* https://git.centos.org (and also the MQTT bus for git push notifications)
* https://cbs.centos.org (and also non public signing service)
* other small services using OpenID/OpenIDC for authentication
(https://blog.centos.org, some jenkins instances used by QA team, etc)
As said, we have already staged all changes to support new auth in our
ansible roles.
When we'll have rolled out these changes, your existing TLS certificate
that you use to authenticate with for cbs.centos.org *will not* work
anymore (important)
That means that you'll have to retrieve a new TLS cert, signed by the
IPA CA cert.
How to do that ? I'll see about how porting this to know repository, but
for now, there is a copr repo that you can use :
https://copr.fedorainfracloud.org/coprs/arrfab/fasjson-client/
IMPORTANT : do *not* use this pkg now, or do this from another
workstation/vm/account/whatever : the new 'centos-cert' util would
replace your currently working TLS cert (from ACO) . (Well, as fasjson
for prod *isn't* deployed yet, that would not work at all, but it would
when deployed
If you have questions, feel free to ask in this thread, or join
#fedora-aaa on Freenode.
--
Fabian Arrotin
The CentOS Project | https://www.centos.org
gpg key: 17F3B7A1 | twitter: @arrfab
_______________________________________________
CentOS-devel mailing list
CentOS-devel(a)centos.org
https://lists.centos.org/mailman/listinfo/centos-devel
Hi folks,
We are planning to update all the plugins (compatible) installed on
ci.centos.org jenkins instance.
I will do it tomorrow morning (Dec 18th) at 9am UTC. I will start
preparing for shutdown then and all the jobs triggered post that will
go in the queue until the instance has updated the plugin and
restarted.
I would also like to open a question to you all who are using OCP4
cluster. How would you want to manage Jenkins update for your
namespace? There are multiple ways from them being auto updated
whenever we update the cluster (this is by default that we use),
time'd/self trigger updates, or updates on change in tags.
I am interested in hearing your thoughts.
Thank you
--
Vipul Siddharth
He/His/Him
Fedora | CentOS CI Infrastructure Team
Due to some network switches upgrade in the DC hosting some community
projects (including but not limited to CentOS), we'll have a large
majority of our infra not reachable.
Migration is scheduled for """"Tuesday November 10th, 2:00 am UTC time"""".
You can convert to local time with $(date -d '2020-11-10 14:00 UTC')
We unfortunately can't announce/give you any expected downtime as it can
last for several hours (info I received through invite) but we'll try to
restore all services/connectivity as soon as possible.
Impacted services in that DC :
- *all*
Non impacted services (easier to just mention short list of things not
in that DC, so items not listed below *will* be down) :
- https://www.centos.org
- https://forums.centos.org
- https://lists.centos.org
- mirrorlist.centos.org
--
Fabian Arrotin
The CentOS Project | https://www.centos.org
gpg key: 17F3B7A1 | twitter: @arrfab