Manual patching Process on Solaris 8 and 9 versions
----------------------------------------------------------------
----------------------------------------------------------------
Verify your disk are okay
# metastat | grep State:
# vxdisk list
* DO NOT PATCH IF YOU HAVE OS DISK ISSUES
- RESOLVE ANY ISSUES BEFORE PATCHING
Check sendmail and make a note if it is running or not
# /usr/bin/ps -ef | grep send
Check and make sure root can run at
# /usr/bin/grep root /etc/cron.d/at.allow
Make some system saves (we do not always need to refer back to these)
# /usr/bin/mkdir /staging/SAVES
# /usr/bin/ps -ef | grep send > /staging/SAVES/sendmail_ps.out
# /usr/bin/uname -a > /staging/SAVES/uname_a.out
# /usr/bin/netstat -rn > /staging/SAVES/netstat_rn.out
# /usr/sbin/vxdisk list > /staging/SAVES/vxdisk_list.out
# /usr/sbin/metastat > /staging/SAVES/metastat.out
# /usr/bin/iostat -En > /staging/SAVES/iostat_En.out
# /usr/bin/iostat -en > /staging/SAVES/iostat_en.out
# /usr/bin/cp -p /etc/mail/sendmail.cf /staging/SAVES/sendmail.cf
# /usr/bin/cp -p /etc/mail/submit.cf /staging/SAVES/submit.cf
# /usr/sbin/mount > /staging/SAVES/mount.out
# cd /etc; tar cf - init.d rc2.d rc3.d | gzip -c > /staging/SAVES/rc_scripts.tar.gz
###################################################
Solaris 8/9
The steps for installing are as follows:
1) Drop the server into single user mode
a) If there are no firmware updates: (init 0, boot –s) – note that running ‘init s’ does not guarantee a quiet system
b) If there is firmware updates staged: Go to the ‘tx flash only.rtf’ doc
2) Mount all the file systems
note, do *NOT* run MOUNTALL, I have seen cases where this brings
the entire system back up to multi user mode.
RUN THE FOLLOWING COMMANDS TO GENERATE YOUR MOUNT COMMANDS:
# /usr/bin/cat /etc/vfstab | \
/usr/bin/tr -s ' ' '\11' | \
/usr/bin/cut -f3,6 | \
/usr/bin/tr '\11' ':' | \
/usr/bin/grep ':yes' | \
/usr/bin/cut -f1 -d: | \
/usr/bin/grep -v '/global' | \
/usr/bin/sed -e "s/^/mount /"
...This will generate the mount commands, but it will not mount anything, that is our job
3) Initiate the patch scripts:
# /usr/bin/chmod 0755 /staging/manual_patching
# /staging/manual_patching &
# /staging/manual_monitor
2) Once the server reboots, verify the patching is running by logging
in through the console as root and execute:
Check the time stamp - should be current
# /ust/bin/ls -al /staging/output.other_patches
Check the progress of the patching
# /ust/bin/tail /staging/output.other_patches
* to monitor patching progress run '/staging/monitor'
3) Once the server is back online check the disk and messages file.
4) Restore sendmail:
# /usr/bin/cp /staging/SAVES/sendmail.cf /etc/mail/sendmail.cf
# /usr/bin/cp /staging/SAVES/submit.cf /etc/mail/submit.cf
5) If sendmail was *online* previously, we want to restart it
(check '/staging/SAVES/sendmail_ps.out' ):
# /etc/init.d/sendmail stop
# /etc/init.d/sendmail start
6) If sendmail was NOT running previously we need to stop and disable sendmail
(check '/staging/SAVES/sendmail_ps.out' ):
# /etc/init.d/sendmail stop
# /usr/bin/mv -f /etc/rc2.d/S88sendmail /etc/rc2.d/DISABLED_S88sendmail
7) Verify the new kernel revision level
# uname -v
No comments:
Post a Comment