CentOS Pulse #1001 - 02 March 2010
Subscribing to this newsletter? We have set up a read-only centos-newsletter mailing list where we announce new CentOS Pulse releases to subscribers with a summary and a link. You can subscribe here! Further there is a Newsletter RSS feed available under http://feeds.feedburner.com/feedburner/PZYQ |
Contents
- Foreword
-
Interview
- How many instances of CentOS do you roughly run? (Kernels only, so Xen instances count as one.)
- What do you like the most about CentOS?
- What really drives you up the wall?
- How long have you been using CentOS in your organization?
- How do you manage packages/installs maintenance?
- What software do you use? (NFS, Xen, ldap, Apache, etc...)
- What is really special about your installation?
- How many people do you have to administer this setup?
- What is the biggest problem you have ever encountered while running CentOS?
- What would you like to see in the future?
- What is the the next big thing you are going to approach while using CentOS? (Using ksplice, for example?)
- Community Threads
- Tip Of The Newsletter
- FOSDEM Report
- Jokes and Funny Stuff
- CentOS Errata
- CentOS in the Spotlight
- Upcoming Events
- Contributing to this newsletter
1. Foreword
After some delay we present another Newsletter. In this issue we have a very interesting interview on the usage of CentOS at University College London, a report on FOSDEM 2010 (where nearly all of the main CentOS personnel showed up) and, of course, the usual categories like community, jokes and updates.
Enjoy the read.
The Newsletter Team.
2. Interview
This month we were lucky enough to interview one of the SysAdmins at UCL (University College, London). We are always looking out for interesting people to interview, so if you want to contribute, please just speak up on the centos-docs mailing list.
2.1. How many instances of CentOS do you roughly run? (Kernels only, so Xen instances count as one.)
150 Fully Managed, 50 Half Managed and an approximate 200 - 250 Rocks cluster. (Based on CentOS-4 & -5.)
We have a concept of fully managed machines where all machine management is looked after centrally.
Lab machines and most desktops - Half-managed machines, where they access a more restricted set of resources but the user has sudo access and can install packages. However, we still look after patch management. These are used for specific project machines and by certain types of user.
Cluster machines - We have a reasonable sized computational cluster that runs Rocks, which is itself based on CentOS.
2.2. What do you like the most about CentOS?
It's not Windows! In three words, Stability, Reliability & Cost. We made the move from Red Hat when they split RHEL-5 into Server or Desktop installations. We had been considering, for a while, why we payed for services from Red Hat that we never used. CentOS seemed to be the smart choice and early tests showed that their were no problems with the existing management systems, so we slowly rolled CentOS out as we upgraded machines from RHEL-3 or -4.
2.3. What really drives you up the wall?
Same as all SysAdmins - the users. Seriously, trying to explain to users they don't need root access to install x bit of software and they don't need a particular distribution. Also the licensing models of some commercial software products ... a well know bit of computational software costs us far too many man hours.
2.4. How long have you been using CentOS in your organization?
Three years now (and Red Hat for 10-12 years).
2.5. How do you manage packages/installs maintenance?
We have bespoke software to manage all the Unix machines -- from Solaris, Irix to Linux. This was written in-house 20+ years ago. It is very flexible and does exactly what it needs to do. However we wish to change to an open-source machine management solution. It is finding the time test such packages, choose one and then roll it out.
2.6. What software do you use? (NFS, Xen, ldap, Apache, etc...)
We use CentOS for most user desktop machines and all of our virtual server infrastructure.
CentOS-ds/typo3/trac/mediawiki/twiki/alfresco/tomcat/drupal/apache/sql/condor - no computer lab cycles are left spare.
We also have lots of custom written software.
With the cluster, we run lustre and nfs on CentOS-5 boxes. We also have the whole Rocks sge scheduler and login nodes running on virtualised servers.
2.7. What is really special about your installation?
Being a computer science research department we have to be very flexible and responsive to our users and often have to run bleeding edge software/hardware. We have many examples of users writing their own software which then gets run as a service for others. We are traditionally more open than in a business environment. We run an extremely diverse environment ... everything from Sun kit running SunOS to fully virtualised CentOS blades to the latest clustered filesystems.
2.8. How many people do you have to administer this setup?
Two full-time and one part-time.
2.9. What is the biggest problem you have ever encountered while running CentOS?
Biggest headache is supporting new hardware, drivers, etc, especially on laptops. (We usually end up recommending Fedora on the latest laptops.) That is the price you pay for a stable, long-life distribution.
2.10. What would you like to see in the future?
We would love a decent virtual machine manager to be available in CentOS, with more of the feature sets of Xencenter/VCentre. We would also like to see ZFS ported to CentOS.
2.11. What is the the next big thing you are going to approach while using CentOS? (Using ksplice, for example?)
Our next major change will be an evaluation of kvm instead of Xen and the test & roll out of CentOS-6. The other big thing at the moment is the explosion in storage ... In the cluster space we will be doing quite a bit of work on the faster/higher bandwidth storage subsystems.
For anyone interested, this is what the department is currently up to ...
http://www.cs.ucl.ac.uk
http://www.cs.ucl.ac.uk/research
3. Community Threads
This section of the Newsletter dives into interesting contributions inside the CentOS community. We make the distinction between contributed documentation on the Wiki, support-related contributions on the mailing lists & fora and development-related topics on centos-devel. We need hints from the community to find these interesting threads as well as to contribute on reporting different viewpoints, summaries and insights.
3.1. CentOS L10n
In our last newsletter we mentioned Timothy Lee's effort in the CentOS wiki Chinese translation. Translating parts of the CentOS project to different languages may be useful for non-English speaking users. The localization process ("L10n", where the 10 is being a way of replacing the middle 10 letters of the word) contains many aspects, starting off with the translation of a specific component and leading into change tracking. This guarantees that translated content is always up to date.
If you are interested in taking part, you should first decide whether you want to translate wiki pages, anaconda slides or just assist on international fora or lists. Afterwards, please introduce yourself on the -docs ML and join the translation teams. (One may possibly already exist for your language.)
Please take a look at the Contribute page for further information.
4. Tip Of The Newsletter
If you have a good tip or know of a really good program that you would like to share, please send us an email.
4.1. Obtaining accurate server time is really causing me difficulties
With Xen instances not being in total sync and time being vital for so many things this is a really important issue. You don't want to be relying on servers that have wrong time-stamps especially when you are comparing logs from different machines to find an error.
Question: I have a problem with CentOS time setting. I have two time servers listed and have Enable Network Time Protocol selected but the system will not update itself. I have to go into Date & Time setting, click on the Network Time Protocol tab and then click OK for it to update. Any ideas?
- Answer: The GUI is not really well suited to resolving this issue.
- As root, in a console:
/sbin/service ntpd stop ; /usr/sbin/ntpdate -s tick.usno.navy.mil && /sbin/hwclock -w ; /sbin/service ntpd start
- If that does not work, you have issues out of scope of the local time management application.
- As root, in a console:
It might be worth finding a time server nearer to you -- http://www.pool.ntp.org/ is quite good for this.
Thank you AkemiYagi for this excellent tip.
5. FOSDEM Report
As FOSDEM is now over, it's time to reflect. This year I travelled with my colleague Sandro, who is a Fedora Developer. We arrived on Friday evening and had a nice dinner in 'The Drug Opera'. The mousse au chocolat was absolutely wonderful. Later on we went to the Friday beer event, which was very crowded. I met Andreas, Ralph & a few others and managed to get ONE drink (which took about half an hour).
Next day, when we were setting up the booth, Didi was a bit late so we had to wait for the Promo material to arrive. However we were not the only ones who were waiting -- all the stickers were gone within a few hours and we sold a lot of T-Shirts, too.
In the evening I gave a talk about Spacewalk, an enterprise lifecycle management system for CentOS, RHEL and Fedora. The audience was pleasant and interested.
On Sunday we were quite lazy but shot some cute photos of Didi in 'Other Linux Distro' shirts. (Now we know what the CentOS metal bling is for.)
Overall it was a great weekend. Even the Internet connection at the FOSDEM event occasionally worked! Hopefully, we will get some new stock for the promo box before the next event.
On Behalf of the CentOS family,
Photo (c) by FabianArrotin and Chris Geldenhuis
6. Jokes and Funny Stuff
6.1. WAIT
Knock Knock!
Who’s there?
.....
WHO?
.... long wait
Java
6.2. sudo su
6.3. CentOS booth
7. CentOS Errata
This section highlights the most severe security updates for each supported CentOS release, whilst providing a summary and short links to the reference of the security issue.
7.1. CentOS-3
2009:1531 seamonkey security update for CVE-2009-1563, 2009-3274, 2009-3375, 2009-3376, 2009-3380
2010:0029 krb5 security update for CVE-2009-4212
2010:0113 seamonkey security update for CVE-2009-1571, 2010-0159
7.2. CentOS-4
2009:1530 firefox security update for CVE-2009-1563, 2009-3274, 2009-3370, 2009-3272, 2009-3273, 2009-3274, 2009-3375, 2009-3376, 2009-3380, 2009-3382
2009:1531 seamonkey security update for CVE-2009-1563, 2009-3274, 2009-3375, 2009-3376, 2009-3380
2009:1601 kdelibs security update for CVE-2009-0689
2009:1673 seamonkey security update for CVE-2009-3979, 2009-3983, 2009-3984
2009:1674 firefox security update for CVE-2009-3979, 2009-3981, 2009-3983, 2009-3984, 2009-3985, 2009-3986
2010:0029 krb5 security update for CVE-2009-4212
2010:0094 HelixPlayer security update for CVE-2009-4242, 2009-4245, 2009-4247, 2009-4248, 2009-4257, 2010-0416, 2010-0417
2010:0112 firefox security update for CVE-2009-1571, 2009-3988, 2010-0159, 2010-0160, 2010-0162
2010:0113 seamonkey security update for CVE-2009-1571, 2010-0159
7.3. CentOS-5
2009:1601 kdelibs security update for CVE-2009-0689
2009:1674 firefox security update for CVE-2009-3979, 2009-3981, 2009-3983, 2009-3984, 2009-3985, 2009-3986
2010:0029 krb5 security update for CVE-2009-4212
2010:0112 firefox security update for CVE-2009-1571, 2009-3988, 2010-0159, 2010-0160, 2010-0162
8. CentOS in the Spotlight
The following articles mention CentOS and are a good resource to understand how the media (and public) view CentOS.
Active Systems has released the first public beta of OpenNode, a CentOS based distro that is designed to make it easier to run virtual machines using KVM or OpenVZ.
Even though not directly related to us, they are still worth a read to see what we are going to get.
'ClearOS is based on CentOS ...'
9. Upcoming Events
The CentOS Promo SIG organizes CentOS presence (booths, presentations, etc) at various conferences and tradeshows. Here we highlight upcoming events. If you are interested in helping, join the Promo SIG.
2010-03-13 → 2010-03-14: Chemnitzer Linuxtage 2010, Chemnitz, Germany -- official site
2010-06-09 → 2010-06-12: LinuxTag 2010, Berlin, Germany -- official site
2010-08-21 → 2010-08-22: FrOSCon 2010, St. Augustin, Germany -- official site
2010-09-17 → 2010-09-18: FrOSCamp 2010, Zurich, Switzerland -- official site
2010-11-06 → 2010-11-07: T-DOSE 2010, Eindhoven, the Netherlands -- official site
10. Contributing to this newsletter
We are always on the look-out for people who are interested to help:
- report on CentOS community activity
- maintain a section in the newsletter
- write an article (or hint) on an interesting topic or person
If you would like to appear in the newsletter, you will have to have contributed positively to the CentOS community and then, hopefully, get noticed by one of our reporters.
We have a special page with more information about contributing!