Bryce T. Pier wrote:

> So I'm curious what other people are doing on the Linux platform.

If you are able to run in a virtualised environment (even in a "one on one" one 
virtual to one physical environment) then you may have snap-shotting.

And on (Open)Solaris, I believe you can snapshot the ZFS root, and roll back as 
required.

Be careful about rolling back.  It's obvious, but I saw someone take a snapshot 
of a live system, whilst it was running (RedHat as it happens).  After the 
upgrade, VMWare tools were reinstalled.  This trashed the fstab (it added its 
own private stuff to a backup of the fstab it had taken when first installed, 
and therefore lost all subsequent updates).  The system was rolled back.  This 
reinstated the server (including memory state) to the live state it was in 
prior 
to patching.  Unfortunately the Oracle database was on an NFS server, which 
wasn't rolled back and the in-core version of Oracle was out of kilter with the 
NFS databases. The result was an Oracle database beyond repair.

Other things that have gone wrong?

I have seen SLES/Yast get upset and create a mess.

A recent SLES patch overwrote Exim's sendmail binary on our majordomo server 
with a copy of Postfix (possibly our fault).

I have also seen an application (Groundwork on SLES) which installs its private 
library path which contains duplicates of system libraries (e.g. libcurses, off 
the top of my head) into /etc/ld.so.conf . When you upgrade the kernel, the 
SLES 
initramfs image ends up with calls to libraries in /usr/local which aren't 
there 
at boot time.  System un-bootable.  A pain to recover.

Touch wood, I've never had a problem with RedHat or Ubuntu, though the 
roll-back 
issue could strike any distro.

Good luck!
-- 
Jonathan
_______________________________________________
Discuss mailing list
[email protected]
http://lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to