RHEL6 and Solaris 10 server - Patching

What is a Patch ?

- A patch is an update to particular software or software bundle.
- Patch is released to fix the issue, or small security issue, so it can be a small file or can be collection of many files which may come with new features.
- So, patches can be for a specific software components, such as the kernel, or a service, as httpd.
- Patches are released to public after testing, and verifying and deploying.
- As an administrator, you have to check vendor site and read articles related to their product and their releases.
- Upon the release, test the patch on your sandbox environment and than, on your DEV env, test env, UAT and finally PROD.
- Make sure wait for a week once you applied the patch to see if anything breaks.
- Some patches are release daily, weekly, monthly and even quarterly.

patches are released to fix,
- Bug fix
- security issue fix
- software update/patch
- kernel patch


Solaris 10 / 11 patching
- Oracle releases patches on monthly or quarterly basis, just download and apply it.
- Read the ReadMe file which gives your detail process on how to install it.
- Solaris 11 comes with concept of publishes which makes life way easier than on solaris 10.

Redhat
- Once you register your system, you can simply download the patch from redhat and apply on your system.
- Yum is a tool and using yum repo is easy solution to apply patch on RHEL based systems.

===================================================

Patching Instruction:

Redhat 6
Patching Redhat and Solaris servers

1. Create a screen session
# screen

2. Register the host to the correct Satellite chanel
# rhnreg_ks --activationkey=1-123232343434dfsgfgdfgg  --serverUrl https://yourhost.local/abc --force

3. Update your system and reboot.
# yum clean all
# yum update –y

4. Once the patching is done, exit out of screen
# Init 6 (or reboot)

Solaris 10 (zfs)
1. If patch cluster is available over nfs, mount your nfs share. If not copy.
# mount nfsserver:/nfs/share /mnt

2. Delete old boot env if you have one
# ludelete old_BE

3. Create a new boot environment.
# lucreate –n newBE.$(date)

4. Go to patch location and install the prereq patches.
# cd /mnt/10_Recommended
# ./installcluster --apply-prereq --s10patchset
# ./installcluster  -B newBE.$(date) --s10patchset
# luactivate newBE.$(date)
# init 6

Note: Do not use reboot/halt command

Solaris 10 (ufs)
1. If patch cluster is available over nfs, mount your nfs share. If not copy.
# mount nfsserver:/nfs/share /mnt

2. Break the mirror if you have any.
# metadetach - for SVM mirrored volumes. [ meta device ]
3. Delete old BE if you have
# ludelete old_BE
# ludelete c0t0d0s0

# lucreate –m /:/dev/dsk/<disk name>:ufs –n <disk name>
# lucreate –m /:/dev/dsk/c0t0d0s0:ufs –n c0t0d0s0

# cd /mnt/10_Recommended
# ./installcluster --apply-prereq --s10patchset
# ./installcluster  -B <disk name>  --s10patchset
# ./installcluster -B c0t0d0s0 --s10patchset
# luactivate <disk name>
# luactivate c0t0d0s0
# init 6

Note: Do not use reboot/halt command, pls use init command.