FYI, to share my perspective, it's biased since my work is confluent.
Generally, confluent will tend to 'rhyme with' xCAT a bit more. I was a big
part of xCAT 2's design and that carried over as a baseline for confluent. It
is different, but in ways that I think we got right, it'll be similar.
The noderange syntax and behavior is very similar (confluent adds pagination
and a few more syntax things). The concept of groups and inheritance with
formulaic values is similar (though using syntax like {node}-bmc instead of
|$|-bmc| and 192.168.{n1}.{n2} instead of |192.168.($1).($2)|.)
The OS profiles are where things really start to deviate. In confluent, a
profile is a pure filesystem thing. Adding a postscript is a matter of
symlinking or adding a file to
/var/lib/confluent/public/os/{profile}/scripts/post.d/, rather than modifying a
table. We continue to implement diskless, cloning, and scripted install of
Autoyast/subiquity/anaconda.
It doesn't require a DHCP server, but it can also more easily coexist with
uncoordinated DHCP servers.
Everything is built upon an API that is always mapped to REST, and we are
working towards releasing a UI to actually take advantage of it for those that
want a UI for usage or for some big display. There is a WebUI today, but it is
kind of bare bones, and a rewritten UI is coming that covers more of the
functionality (though CLI and API are going to remain the only way to get to
all functionality).
It's a bit more concerned about security, which sounds good, though sometimes
it can be a little picky in weird ways. Like requiring a setting global to
opt-in to PXE support even though pretty much everyone would want that setting.
Our security team has been pretty adamant about explicit opt in, so we do it,
but it's not too bad. SSH is now handled with certificate authorities instead
of passing/sharing private host keys, and host based authentication for node to
node ssh (a bit more flexible than passing user key around and 'zoning').
HTTPS is also well enabled, with nodes all having the appropriate certificate
authority provisioned. Diskless images are encrypted and node TPMs are used
for diskless usage so that node credentials are better protected.
Another complication is that there's several more ways to start. You can PXE
boot and collect mac addresses, but you can also do BMC driven discovery
instead, or just add BMCs manually and run 'nodeinventory nodes -s' to get
there. Which is nice, but requires better documentation so you don't end up
wasting time with an approach you don't like.
The biggest problems is that we haven't made ourselves as readily reachable
(mostly wondering where people want us reachable, Matrix or a forum or...) and
we have struggled to make navigable documentation (which is part lack of
sufficient effort, and frankly not very good at retaining a fresh perspective
for new users). Also, while it does work with various vendor equipment, I'm the
lead and work for Lenovo, and that may be a concern for some folks who don't
want hardware vendors writing their management software, or at least don't want
Lenovo doing it.
Mostly I hear about alternatives that are about OS deployment, so not as many
as concerned with deep BMC operation. So I think that's something that I think
confluent is uniquely similar to xCAT (though more so, e.g. nodeconfig can
modify UEFI settings on Lenovo, Dell, HPE, and others). Other than that, I
can't give a good objective opinion, since of course I'll like confluent more
since it's exactly what I want it to be.
I think I'll pivot to more documentation and videos after we get the WebUI
rewrite out, and use that as one of the options for easier guided onboarding
for people to get started. I'll maintain CLI-only of course, as that's a bit
more my style, but I'm hoping the GUI can alleviate the rough edges for those
that appreciate it.
________________________________
From: Noah, Stuart via xCAT-user <[email protected]>
Sent: Monday, October 21, 2024 5:11 PM
To: xCAT Users Mailing list <[email protected]>
Cc: Noah, Stuart <[email protected]>
Subject: Re: [xcat-user] [External] Re: xCAT Consortium Update
Hi Brian,
What is your opinion of Confluent?
The message threads lately seem to indicate that that the direction is to
migrate folks to that rather than continue releases of Xcat…
We have our new cluster on Bright, but that is a cash hog… and overkill for
the development environment that we are currently redesigning.
Thanks,
Stu
From: Russell Jones <[email protected]>
Sent: Saturday, October 12, 2024 7:40 AM
To: xCAT Users Mailing list <[email protected]>
Subject: [External] Re: [xcat-user] xCAT Consortium Update
For what it's worth all, I headed up migrating the company I work for to
Warewulf. There's a bit of a learning curve because it does things quite
differently, but the documentation is decent and the developers are very active
in their
For what it's worth all, I headed up migrating the company I work for to
Warewulf. There's a bit of a learning curve because it does things quite
differently, but the documentation is decent and the developers are very active
in their Slack channel and receptive to helping when you get stuck.... and
there were several things I got stuck on :-) One is we needed to be able to
support multiple networks from a single manager just like xCAT (warewulf
doesn't have that capability out of the box). The developers helped, and it was
totally doable with some minor configuration file edits.
Another "major" thing is you are on your own for writing your own systemd
script to launch postscripts after a node boots. Warewulf does not have a
built-in capability for postscripts. Again, totally doable.
I would recommend moving to Warewulf.
On Wed, Oct 9, 2024 at 1:20 PM Glick, Bill
<[email protected]<mailto:[email protected]>> wrote:
Thank you, Markus.
I want to apologize for the tone of my original email. You had given an update
about this on May 10,
2024<https://urldefense.com/v3/__https:/sourceforge.net/p/xcat/mailman/message/58770579/__;!!KOmnBZxC8_2BBQ!3BDzbG93wbVijU0gAjAOFnVSIVYSX71pqAYb4UgJAciajsalv8A98jWYpm4rigz6kC73dc-lQ6WphbxwCFyB$>,
but I was having trouble locating that when I sent this latest enquiry.
I have been using the last development
release<https://urldefense.com/v3/__http:/xcat.org/files/xcat/repos/yum/devel/core-snap/__;!!DZ3fjg!4r59aVCvn4jsWNCCrHAh6xyrgu5mx6BI9J0SxQ-d0TGtijvuu86F-dHJdhhvulngQh7UVfY6X1ZJUe-x7E-_X1DEjgU$>
for the last week. It is working well for us. Thanks!
- Bill Glick
From: Markus Hilger
<[email protected]<mailto:[email protected]>>
Date: Tuesday, October 8, 2024 at 10:56 AM
To: Glick, Bill <[email protected]<mailto:[email protected]>>, xCAT Users
Mailing list
<[email protected]<mailto:[email protected]>>
Subject: AW: xCAT Consortium Update
Hi Bill,
yes, we are still here :)
However, as we are all only following the project on the side, things are
sometimes progressing a little slower than we would like.
The current plan is to release
2.17<https://urldefense.com/v3/__https:/github.com/xcat2/xcat-core/milestone/53__;!!DZ3fjg!4r59aVCvn4jsWNCCrHAh6xyrgu5mx6BI9J0SxQ-d0TGtijvuu86F-dHJdhhvulngQh7UVfY6X1ZJUe-x7E-_I0DJmgo$>
before Supercomputing this year. The codebase is ready, and the last
development
release<https://urldefense.com/v3/__http:/xcat.org/files/xcat/repos/yum/devel/core-snap/__;!!DZ3fjg!4r59aVCvn4jsWNCCrHAh6xyrgu5mx6BI9J0SxQ-d0TGtijvuu86F-dHJdhhvulngQh7UVfY6X1ZJUe-x7E-_X1DEjgU$>
is what 2.17 will be.
So, anyone is welcome to use it already. You don't have to wait for an
"official release".
We might want to hold another BoF session at Supercomputing. But we are still
finalizing the details.
Nevertheless, I would like to mention again that xCAT itself has no long-term
future. We do not recommend using xCAT for new clusters that may want to
upgrade to newer operating systems such as RHEL 10.
The project is only in maintenance mode. We will provide a small fix here and
there for existing installations but there will be no major new development or
regular releases.
We are happy to merge pull requests though.
For new deployments we recommend using
confluent<https://urldefense.com/v3/__https:/github.com/xcat2/confluent__;!!DZ3fjg!4r59aVCvn4jsWNCCrHAh6xyrgu5mx6BI9J0SxQ-d0TGtijvuu86F-dHJdhhvulngQh7UVfY6X1ZJUe-x7E-_IdZDBVg$>
instead.
Mit freundlichen Grüßen / Kind regards
Markus Hilger
HPC Engineer
MEGWARE Computer Vertrieb und Service GmbH
Tel: +49 3722 528-47
Nordstraße 19
[email protected]<mailto:[email protected]>
09247 Chemnitz-Röhrsdorf, Germany
www.megware.com<https://urldefense.com/v3/__http:/www.megware.com/__;!!DZ3fjg!4r59aVCvn4jsWNCCrHAh6xyrgu5mx6BI9J0SxQ-d0TGtijvuu86F-dHJdhhvulngQh7UVfY6X1ZJUe-x7E-_gOWqpSY$>
Geschäftsführer: André Singer, Dr. Axel Auweter
Amtsgericht: Chemnitz HRB 584
________________________________
Von: Glick, Bill <[email protected]<mailto:[email protected]>>
Gesendet: Mittwoch, 2. Oktober 2024 18:13
An: xCAT Users Mailing list
<[email protected]<mailto:[email protected]>>;
Markus Hilger <[email protected]<mailto:[email protected]>>
Betreff: Re: xCAT Consortium Update
Any news or updates regarding the xCAT Consortium?
Did discussion move to some new place? Or are we to presume the xCAT Consortium
is defunct?
- Bill Glick
From: Glick, Bill <[email protected]<mailto:[email protected]>>
Date: Thursday, May 2, 2024 at 10:33 AM
To: xCAT Users Mailing list
<[email protected]<mailto:[email protected]>>
Subject: Re: [xcat-user] xCAT Consortium Update
How is the xCAT Consortium effort going?
I had the impression we’d hear approximately monthly updates about its
progress, but I haven’t seen any updates since this message in December.
- Bill Glick
From: Markus Hilger
<[email protected]<mailto:[email protected]>>
Date: Monday, December 11, 2023 at 11:07 AM
To: xCAT Users Mailing list
<[email protected]<mailto:[email protected]>>
Subject: [xcat-user] xCAT Consortium Update
Dear xCAT community,
it is one month ago, since we met at Supercomputing ’23 to announce that a
consortium led by RedLine, OCF and MEGWARE is willing to take over the
ownership of the current xCAT project to provide continuity for existing xCAT 2
based clusters. And we are still thrilled having seen so many of you during
this event who have shown their interest and support, both in presence and
online.
On the technical side, we have started to dive into the IBM-internal
integration setup for compiling and testing xCAT releases. Our goal is to
replicate and extend this infrastructure at RedLine, OCF and MEGWARE to provide
even wider test coverage on more hardware infrastructures for future releases.
At the same time, we continue to investigate the various legal aspects of the
transition.
During our SC’23 meeting, we have also announced that in addition to our
efforts surrounding xCAT 2, we would work with LENOVO towards making Confluent
a suitable, community backed, xCAT 2 replacement for future cluster systems.
For this reason, we invited LENOVO to become a full member of our consortium
from the beginning. Understandably, this plan is a significant step for the
Confluent team that requires thorough preparation and careful consideration and
we will share further updates as this effort takes shape.
We would also like to take this opportunity to thank the team at IBM, namely
Nathan A Besaw, for their continued support during the transition.
Mit freundlichen Grüßen / Kind regards
Markus Hilger
HPC Engineer
MEGWARE Computer Vertrieb und Service GmbH
Tel: +49 3722 528-47
Nordstraße 19
[email protected]<mailto:[email protected]>
09247 Chemnitz-Röhrsdorf, Germany
www.megware.com<https://urldefense.com/v3/__http:/www.megware.com/__;!!DZ3fjg!408lxur06Jk8-bQWQGKw6CU6hJUpB8yNM9CuibvOFzrB8wxMVavWBFshC-oqQggakVDDlr1jpM49BBSszZMCWMo_ObU$>
Geschäftsführer: André Singer, Axel Auweter
Amtsgericht: Chemnitz HRB 584
_______________________________________________
xCAT-user mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/xcat-user<https://urldefense.com/v3/__https:/lists.sourceforge.net/lists/listinfo/xcat-user__;!!KOmnBZxC8_2BBQ!3BDzbG93wbVijU0gAjAOFnVSIVYSX71pqAYb4UgJAciajsalv8A98jWYpm4rigz6kC73dc-lQ6WphXb-0FnO$>
IMPORTANT WARNING: This message is intended for the use of the person or entity
to which it is addressed and may contain information that is privileged and
confidential, the disclosure of which is governed by applicable law. If the
reader of this message is not the intended recipient, or the employee or agent
responsible for delivering it to the intended recipient, you are hereby
notified that any dissemination, distribution or copying of this information is
strictly prohibited. Thank you for your cooperation.
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user
