Upgrade to CentOS 6.7 from CentOS 6.6/6.5/6.4/6.3/6.2/6.1/6.0 - Comment Page: 1

CentOS 6.6 is released. This is quick guide, howto upgrade to CentOS 6.7 from CentOS 6.6/6.5/6.4/6.3/6.2/6.1/6.0. I have tested this method on multiple CentOS 6 (servers / desktops) and everything looks work fine after upgrade. Upgrade to CentOS 6.7 from CentOS 6.6/6.5/6.4/6.3/6.2/6.1/6.0 1. Change to root User su - ## OR ## sudo -i 2. Backup all important data Backup /etc diretory Backup important logs /var/log Backup web server configs and sites Dump MySQL databases Dump PostgreSQL databases Backup all what you need if something goes wrong [inttf_post_ad1] 3. Check list of packages that are going to be updated yum list...

26 comments on “Upgrade to CentOS 6.7 from CentOS 6.6/6.5/6.4/6.3/6.2/6.1/6.0 - Comment Page: 1

1 2
    1. Excellent instruction.

      Reply
    2. I guess since from today we have CentOS 6.2 realesed , the process of updating from 6.1 to 6.2 will be the same.

      Reply
      • Hi Igor,

        I noticed also that the CentOS 6.2 is released! :)

        The process of updating is exactly same and I updated this guide also…

        I updated three CentOS 6.1 machines to just test this guide and one box I have to run:

        
        yum clean all
        yum update
        

        because yum did not find any updates. It was due to either local cache or the fact that all the mirrors may not be even up-to-date yet.

        Reply
    3. you may add in this

      yum -y install yum-plugin-fastestmirror

      Reply
    4. > “everything looks work fine after upgrade.” (6.2 >> 6.3)

      No, it doesn’t.

      Upgrading to 6.3 wipes out config files right and left, so, if you done something as simple as moving your PHP session.save_path, so you can run apache under a username, your localhost virthosts are fragged. I’m still tracking down what else it has overwritten, as that alone doesn’t get virthosts working again.

      It also trashes rebuilding Broadcom wireless cards. (Granted Broadcom wireless cards are carp.)

      *sigh* Why do I ever try to upgarde once a system is working….

      Best,
      Michael

      Reply
      • Hi Michael,

        Could you just diff your config files from backups?

        Normally upgrade process should keep old configs. I have upgraded several servers from 6.0 to 6.1, 6.1 to 6.2 and 6.2 to 6.3 with this method and no problems at all.

        Reply
    5. Now JR, teasing people isn’t nice!

      Where are these “old configs?” Or, even just point me at the script in the upgrade process that makes them…

      (As there seems to be no real 6.x docs/wiki, my searches for same have proven fruitless.)

      Best,
      Michael

      Reply
      • Hi again Michael,

        Sorry, but first, I thought that you tried to upgrade your system according to this guide.

        More info, howto RPM package config files work:
        RPM, %config, and (noreplace)

        So you might have .rpmsave files (if config was overwritten) or .rpmnew files (if new default config available), so you can search configs with following commands (as root):

        
        find /etc /var -name *.rpmsave
        
        find /etc /var -name *.rpmnew
        

        Then just diff your findings with current configs, like:

        
        diff -u /etc/php.d/apc.ini /etc/php.d/apc.ini.rpmsave
        
        Reply
        • Hi JR,

          Even if I had followed your guide, my Apache children would have still segfaulted. Little buggers just can’t read!

          Although I wish I had followed your instructing, as it would have saved me re-creating a couple config files.

          Here’s the relevant fix for “child pid 2350 exit signal Segmentation fault” after an upgrade (found in /etc/httpd/logs/error_log). https://www.centos.org/modules/newbb/viewtopic.php?viewmode=flat&topic_id=38665&forum=56

          Best,
          Michael

          Reply
          • Hi Michael,

            Upgrades are always risky, something unpredictable could happen. There are solutions to help with this type problems (actually help to avoid this type problems). Simplest is full backup and old working system restore if something goes wrong (this is best method for not critical desktops). Another and even much more elegant way on server side is setup high availability (HA) cluster. I have used normally Heartbeat + DRBD setups for most critical production systems, like databases, file servers etc.

            But yes, I could of course add full backup recommendation to this guide, but I have see that many times it’s easier just skip backup steps and try upgrade first… ;) It’s just like, when you buy something new, you test it first and if you have some (serious) problems then you read manual… :)

            It’s nice to hear that you got your system working and this segfault problem fixed and posted it here. Many thanks!

            Reply
            • One additional note to Apache segfaults. I recommend to backup web server configs and sites in backup step. In fact, that would be a way to solve this issue directly. Because you should have right permissions on your backup(s). :)

              Reply
    6. Great, simple instruction!
      I’ve done my CentOS update (from 6.3 to 6.3) without any problems. Many Thanks! :-)

      Reply
    7. sir after update when i start my laptop i have 4 centos
      an i must choose one before continue or let the system choose one for me
      how to make it just one ?
      thank you

      Reply
      • Hi bidbin,

        So you have multiple kernels listed on grub? It’s just normal, but if you want “see” only one kernel then you have to remove older ones, or at least remove older kernels from grub.conf.

        Could you post output of following commands:

        
        yum list installed kernel\* |sort
        
        uname -a
        
        cat /boot/grub/grub.conf
        
        Reply
    8. Thanks you… I updated my VPS server to CentOs 6.3… Very helpful..!!!

      Reply
    9. I updated my VPS server to CentOs 6.3 without any problems.
      it was very easy way.
      Thanks,

      Reply
    10. I’m so happy to read this. This is the type of manual that needs to be given and not the accidental misinformation that’s at the other phorums. Appreciate your sharing this greatest doc. Great!

      Reply
    11. I am newbie can anyone post steps for backup and restore for following things

      Backup all important data
      Backup /etc diretory
      Backup important logs /var/log
      Backup web server configs and sites
      Dump MySQL databases
      Dump PostgreSQL databases

      Reply
    12. thanks,,,updated to centos6.4 from 6

      Reply
    13. Hi,

      I have 3 centos 6.2 servers which will not go online. Recently i have setup a centos 6.4 server which will go online and I have setup yum repo. I want to upgrade the cent 6.2 to 6.4 is it possible to do with yum update in 6.2, if i point to the 6.4 repo.

      Reply
    14. When I gave update from centos 6.2 to 6.4 it gave the following error.

      Error: Package: e2fsprogs-libs-1.41.12-14.el6_4.2.x86_64 (base)
      Requires: libcom_err = 1.41.12-14.el6_4.2
      Removing: libcom_err-1.41.12-11.el6.x86_64 (@anaconda-CentOS-201112091719.x86_64/6.2)
      libcom_err = 1.41.12-11.el6
      Updated By: libcom_err-1.41.12-14.el6.x86_64 (base)
      libcom_err = 1.41.12-14.el6
      You could try using –skip-broken to work around the problem
      You could try running: rpm -Va –nofiles –nodigest

      then i used yum update –skip-broken and upgraded to 6.4(Test Server). Now i have to carry out in production server I got the same issue. But i want to fix the issue without giving –skip-broken.
      Any inputs will be helpful.

      Reply
      • Hi geek,

        So you tried just yum update command on CentOS 6.2?

        Reply
1 2

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Close