2

Two fold question.

  1. How do I find out when the upgrade from 5.8 to 5.9 happened? Where is this logged?
  2. What is the process of rolling back or downgrading from 5.9 to 5.8?

Background: Supporting a client who is adding a NetApp storage device in the upcoming weeks and the configuration is certified up to 5.8, but not with 5.9. It is then certified with 6.x versions. Since the rest of the servers that will be using the device are all running 5.8 the desired outcome is to roll this one back and find out when it was upgraded (I am semi-new to the project, and the previous admin is long gone).

Kernel versions are the same across the servers, /etc/redhat-release is showing 5.8 for all but one, which is showing

$ more /etc/redhat-release

Red Hat Enterprise Linux Server release 5.9 (Tikanga)

Possibly worth noting, they are running Oracle Enterprise Linux, but for all (most) intents and purposes, they can be treated the same as RHEL.

dwaynehoov
  • 33
  • 1
  • 5

2 Answers2

3

You really shouldn't downgrade at this point. If you're not seeing a system stability or performance issue, keep the system at the 5.9 revision level. There's more risk in attempting to rollback than just dealing with issues as they come on the currently-supported version. Remember, Red Hat strives to keep minor OS revisions consistent and compatible throughout the support lifetime. E.g. a RHEL 5.5 box should run similarly to a 5.9 system.

In terms of logging, check to see if the system had the yum-updatesd daemon is running (with chkconfig or ntsysv). That would run automatic updates on the system and could be the reason why the server was pushed from 5.8 to 5.9.

The other location to examine for update and package installation activity is the /var/log/yum.log - You'll see the timestamps and package actions listed...

Jan 06 19:35:49 Installed: iplike-2.0.2-1.el6.x86_64
Jan 08 12:46:26 Installed: jwhois-4.0-19.el6.x86_64
Jan 13 20:30:55 Updated: selinux-policy-3.7.19-155.el6_3.14.noarch
Jan 13 20:31:20 Updated: selinux-policy-targeted-3.7.19-155.el6_3.14.noarch
Jan 13 20:31:21 Updated: iperf-2.0.5-3.el6.x86_64
ewwhite
  • 194,921
  • 91
  • 434
  • 799
  • 2
    He could also check the modification date of `/etc/redhat-release`. – EEAA Feb 05 '13 at 15:33
  • @EEAA Yah, but where's the fun in that?!? – ewwhite Feb 05 '13 at 15:41
  • Hehe, true. It's just a quick way to check. `yum.log` is a better solution, as it not only delivers the information the OP asked about, but also other useful information. – EEAA Feb 05 '13 at 15:44
  • the first thing that I did was check the timestamp on redhat-release, I wanted to be 100% sure that the date was reliable. Also checked yum.log, nothing there indicating a version upgrade, just some package updates. Also, it looks like yum-updatesd is running on the 5.9 server as well as the 5.8 boxes. I don't think I can reliably say that it was automatically updated. – dwaynehoov Feb 05 '13 at 16:04
  • I think i found the issue. automatic updates are running and there is a discrepancy in the repos listed in the yum.repos.d directory. namely, the server that is at 5.9 had this enabled: [ol5_UEK_latest] name=Latest Unbreakable Enterprise Kernel for Oracle Linux $releasever ($basearch) baseurl=http://public-yum.oracle.com/repo/OracleLinux/OL5/UEK/latest/$basearch/ gpgkey=http://public-yum.oracle.com/RPM-GPG-KEY-oracle-el5 gpgcheck=1 enabled=1 – dwaynehoov Feb 05 '13 at 16:28
  • 1
    @dwaynehoov That's because going from 5.8 to 5.9 is _not_ a version upgrade, it's just a bunch of updates rolled up and given a number (similar to a Windows service pack). – Michael Hampton Feb 05 '13 at 17:29
0

RHEL 5.9 was released on 2013-01-07, so it should be more recently than that.

Take a look at /var/log/yum.log (I'm going off CentOS). There should be some package like redhat-release-notes-5.9 with a timestamp on the installation.

cjc
  • 24,533
  • 2
  • 49
  • 69