CentOS Dojo at LISA14 in Seattle on November 10th, 2014

Join us at the all day (09:00 to 17:00) CentOS Dojo on Monday, November 10th, 2014 at the LISA14 conference in Seattle, Washington.

There will be at least three CentOS board members there (Johnny Hughes, Jim Perrin, and Karsten Wade).

The current topics include:

• CI environment scaling by Dave Nalley
• DevOps Isn’t Just for WebOps: The Guerrilla’s Guide to Cultural Change by Michael Stahnke
• The EPEL Phoenix Saga by Stephen Smoogen
• Docker in the Distro by Jim Perrin
• Managing your users by Matt Simmons

Visit the CentOS Wiki for more information.

CentOS-6.6 is Released

CentOS 6.6 is now released, see the Announcement.

So, the Continuous Release RPMs where released on 21 October (7 days after RHEL 6.6) and the Full Release was done 28 October (14 days after RHEL 6.6).

Enjoy.

Continuous Release Repository RPMs for CentOS-6.6 Released

The CentOS team has released the Continuous Release (CR) Repository RPMs for CentOS-6.6 into their 6.5/cr tree.  See the Release Announcement.

Now a little more about the release process.

1. Red Hat releases a version of Red Hat Enterprise Linux.  In this case the version is Red Hat Enterprise Linux 6.6 (RHEL-6.6), which was released on October 14th, 2014.  With that release by Red Hat comes the source code which RHEL 6.6 is based on.
2. The CentOS team takes that released source code and starts building it for their CentOS release (in this case CentOS-6.6).  This process can not start until the Source Code from Red Hat is available, which in this case was October 14th.
3. At some point, all the Source Code has been built and there are RPMs available, this is normally 1-5 days depending on how many Source RPMs there are to build and how many times the order needs to be changed to get the builds done correctly.
4. After the CentOS team thinks they have a good set of binary RPMs built, they submit them to the QA team (a team of volunteers who do QA for the releases).  This QA process includes the t_functional suite and several knowledgeable system administrators downloading and running tests on the RPMs to validate updating with them works as planned.
5. At this point there are tested RPMs ready, and the CentOS team needs to build an installer tree. This means, take the new RPMs and move them into place in the main tree, remove the older ones RPMs they are replacing, run the build installer to create an installable tree, test that installable tree.  This process can take up to 7 days.
6. Once there is an installable tree, all the ISOs have to be created and tested.  We have to create the ISOs, upload them to the QA process, test them for installs via ISOs (correct sizes, how to split the ISOs, what is on the Live CDs and LiveDVDs to keep them below the max size to fit on media, etc.).  We then also test the installs for UEFI installs, Secure Boot installs (CentOS-7 only), coping to USB Keys and checking the installs that way, etc.  This process can also take up to 7 days.

So, in the process above, we can have vetted binary RPMs ready to go as soon as 5 days after we start, but it may be 14 or more days after that before we have a complete set of ISOs to do a full release.  Thus the reason for the CR Repository.

The CR Repository

The process of building and testing an install tree, then creating and testing several types of ISO sets from that install tree (DVD Installer, Minimum Install ISO, LiveCD, LiveDVD, etc) can take 1-2 weeks after all the RPMs are built and have gone through initial QA testing.

The purpose of the CR repository is to provide quicker access to RPMs for an upcoming CentOS point release while further QA testing is ongoing and the ISO installers are being built and tested.

Updates in the CR for CentOS-6.6

More Information about CR.

CentOS-6.6 Release Notes (Still in progress until the actual CentOS-6.6 release).

Upstream RHEL-6.6 Release Notes and Technical Notes.

Xen4CentOS XSA-108 Security update for CentOS-6

There has been a fair amount of press in the last couple of days concerning Xen update XSA-108, and the fact that Amazon EC2 and Rackspace must reboot after this update:

Rackspace forced to reboot cloud over Xen bug

Amazon Reboots Cloud Servers, Xen Bug Blamed

There are other stories, but those articles cover the main issue.

As KB tweeted, the CentOS-6 Xen4CentOS release is also impacted by this issue and the CentOS team has released CESA-2014:X013 to deal with XSA-108.  There are also 3 other Xen4CentOS updates released:  CESA-2014:X010, CESA-2014:X011, CESA-2014:X012

If you are using Xen4CentOS on CentOS-6, please use yum update to get these security updates ... and like Rackspace and Amazon EC2, you need to reboot your dom0 machine after the updates are applied.

CentOS team at cPanel 2014

The CentOS team will have a booth in the Exhibit Hall for the 2014 cPanel Conference at the Westin Galleria hotel in Houston, Texas from September 30th to October 1st 2014.

CentOS Board members Johnny Hughes (that's me :D) and Jim Perrin will be at the booth whenever the hall is open. 

We are looking forward to lots of discussions and we will have some swag to give out (Tee Shirts .. including the new 10 Year Anniversary tee, Stickers, etc.). We will also be happy to install CentOS on your laptop (or let you do it) ... or if you have a USB key available, we will put a CentOS iso on it for you to use for install later.

If you are going to be at cPanel 2014, come on down and see us!

CentOS Linux 5.11 for x86_64 and i386 is released

The CentOS Linux 5.11 distribution for both the x86_64 and i386 architectures is now released.

If you are running any previous version of CentOS-5 Linux, then you can upgrade simply by using the command:

yum update

ISOs are also available here:

http://isoredirect.centos.org/centos/5/isos/

Please see the Release Announcement and Release Notes for more details.

 

Critical Bash updates for CentOS-5, CentOS-6, and CentOS-7

There is a critical CVE issue in all versions CentOS that has been fixed today.  Please update your CentOS with this command:

yum update

Information about this issue can be found here:

http://red.ht/1msy8D6

and here:

http://red.ht/1uZGljA

The CentOS release announcements are here:

CentOS-5, CentOS-6, CentOS-7

If you have any other questions about this issue, you can ask on the CentOS mailing list here.

CentOS Dojo Cincinnati, OH (4 June 2014)

The CentOS team is happy to announce that we will be having a CentOS Dojo on Wednesday, June 4th, 2014 at the University of Cincinnati's Dyer Hall (Map | Register | Info)

CentOS Dojos are a one day event that bring together people from the CentOS communities to talk about systems administration, best practises, and emerging technologies. The emphasis is to find local speakers and tutors to come together and talk about things that they care about most, and to share stories from their experiences working with CentOS in various scenarios.

Both Jim Perrin and I will be there representing the CentOS Board, and there will be several great speakers (details still being worked out on the exact speaker list).

If you are in the Cincinnati area and use CentOS, please register and come to the event.

See you there !

OpenStack Summit Atlanta 2014

I will be representing the CentOS Project at OpenStack Summit Atlanta 2014 on May 12th to 16th, 2014.

The CentOS Linux distribution and RPMs produced by RDO are one of the major ways that OpenStack is now being deployed worldwide (CentOS was the 2nd most deployed Operating System in the latest OpenStack User Survey).

If you are using, or plan to use, RDO on CentOS (or OpenStack Platform on Red Hat Enterprise Linux) then there are several Red Hat tracks available at the summit that would be very informative.

I am also pretty excited about the Red Hat acquisition of Inktank and I would expect to see some noise about that at the summit as well.  Ceph and GlusterFS make up the current members of our CentOS Storage SIG and I look forward to working with both groups in the future. 

I expect that I will be in the Expo Hall from Monday through Wednesday and I will be somewhere at the Summit on Thursday and Friday as well, so if you see an old dude in a CentOS polo, it's likely me.

CentOS Dojo, São Paulo, Brazil: 5 May 2014

The CentOS Project will be having a Dojo in São Paulo, Brazil on Monday, May 5th, 2014 at Avenida Brigadeiro Faria Lima 3900 - Cj 81 Itaim Bibi Sao Paul (map, register, info).

CentOS Dojos are one day events, organized around the world, that bring together people from the CentOS Communities to talk about systems administration, best practices in Linux-centric activities and emerging technologies of note. The emphasis is to find local speakers to come together and talk about things that they care about most, and to share stories from their experiences working with CentOS in various scenarios.


This event is sponsored By Red Hat and happens two days before FISL 15 in Porto Alegre, Brazil.

The CentOS team will also be at FISL 15 and these are our first ever events in South America, so we are pretty excited.

If you are near  São Paulo or Porto Alegre at the beginning of May, please stop by and visit the team.

CentOS Events in the Next Few Weeks

CentOS has 3 events happening in the next month or so.  Please come out and join us at as many of them as you can.

CentOS Dojo in Santa Clara, California on 31 March 2014.

CentOS Dojo in Denver, Colorado on 10 April 2014.

The CentOS Project at the Red Hat Summit in San Francisco, CA on 14-17 April 2014.

We will have CentOS board members at each of these events.  If you want to find out what is happening the the CentOS world then these are the places to be :)

Createrepo and Parallel Delta RPMs on CentOS-6

Anyone who is using Delta RPMs (AKA - the yum-presto plugin) on CentOS-6 or Fedora knows how much of a PITA it is to generate the drpms.  It takes forever to generate.

Ian Mcleod has posted some new functionality on the Yum Developers mailing list that addresses this issue.

Here is the git repo that contains the new code.

I have created an RPM and SRPM for EL6 that use the tarball from the git repo and add the following functionality:

--delta-workers - The number of worker processes to use for delta
related tasks

--max-concurrent-delta-rpm-size - The maximum total size of uncompressed
rpm payloads that are actively being processed by makedeltarpm at any
given time.

You can get copies of the RPMS from my people.centos.org location for testing. 

The only things that I found with the rpm is:

1.  You will need to also install pyliblzma package from epel to use it.

2.  You should use the same (or lower) number of --delta-workers as compared to --workers 

3.  As discussed in the Yum Developers thread above, you want to make sure to assign a value to --max-concurrent-delta-rpm-size .  The number to assign is bytes ... so 16GB (if you have a 64GB ram machine, and the recommended amount is about 25%) then it would be 16 x 1024 x 1024 x 1024 = 17179869184 (as an example).  That machine also has 24 CPU cores ... so I used 24 for --workers and --delta-workers.

CentOS at Scale 12x

Several members of the CentOS team will be at Scale 12x (http://bit.ly/1hEH6t5) in Los Angeles, California on February 21st to 23nd, 2014.

CentOS will be part of the "Infrastructure.Next at Scale" event ( http://bit.ly/1irKWUm ) event that happens on Friday (21st), and we will have the following talk there:
http://bit.ly/1g2voUB

We will be at the Red Hat Community booth/table on both Saturday and Sunday (22nd and 23rd) with free swag (teeshirts, stickers, etc).

We will have a Birds of a Feather session, details of which will be provided at the table/booth when finalized.

Finally, we will also have a talk titled "CentOS Project Q&A Forum" (http://bit.ly/1bQ5rVS) on Sunday (23rd).

If you are in the Los Angeles area, please stop by and see us at Scale 12x !!!

Kernel-ml, NVIDIA Drivers, and the 3.13.x kernel

ELRepo has a Main Line kernel section to allow using the latest kernel with CentOS-6.  I use this kernel on my laptops and workstations as it usually means better hardware detection and battery life, etc.

If you have the need to use the proprietary NVIDIA drivers with this kernel, as I do on one of my laptops, then you would need to install the latest NVIDIA drivers for each kernel.  No one (that I know of at least) maintains an RPM based module for the kernel-ml series, so downloading the latest version of the drivers from NVIDIA's site is how I handle this situation.

Sometimes with a new kernel tree is begun, in this case the 3.13.x kernel tree, the NVIDIA drivers as posted with not work.  This is the current situation with the 3.13.2 kernel (as installed via ELRepo on CentOS 6.5 ) when using the latest NVIDIA drivers (currently version 338.31).

On the web, I found a good blog entry on how to generally fix the issue, so I thought I would go over how to make it work on CentOS-6.

• Enable the ELRepo repository (if required) using their instructions here.
• Install the ELRepo kernel-ml via their instructions here.
• Download the latest NVIDIA Drivers from the NVIDIA site.
• Change the permissions on the downloaded binary file so it can be executed:

chmod 755 NVIDIA-Linux-x86_64-331.38.run

• Extract the NVIDIA drivers with this command in a terminal:

./NVIDIA-Linux-x86_64-331.38.run -x

• Download the combined (2nd) patch from the above blog and save it (right click, save link as) into the NVIDIA-Linux-x86_64-331.38 directory created when extracting the driver above.
• Apply the patch with this command:

patch -p1 < 02_nvidia_3.13.patch

• Reboot into the new kernel (will get console mode, login as root) ... then go to the  NVIDIA-Linux-x86_64-331.38 that contains the patched code.
• Install the patched driver with this command:

./nvidia-installer

The install should now complete in the normal way, and you should be able to boot into the NVIDIA driver on the 3.13.x kernel-ml series from ELRepo on CentOS-6.5