Re: Remote administration of a 2-node sun cluster running Solaris 5.8
oracle.trainer_at_gmail.com
Date: 01/31/05
- Next message: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Previous message: mario: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- In reply to: mario: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Next in thread: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Reply: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 30 Jan 2005 19:31:14 -0800
I happened to be browsing the questions in this group and I do have to
say that there are no relevant answers that have been posted by Sean
and Kevin. Mark's replies do appear quite stuck up and cryptic to me,
JMO
Mario, let me try to get you started...
You can administer a Sun system entirely from a serial (character)
session, including send a "break" to get to the Openboot prompt; The
Java webstart GUI doesn't give you any more functionality.
As regards patch CDs, the patches can be copied across the network onto
the server before installation, if you don't have anyone "local" who
you trust to put a CD in the server.
And you should have a look at WebMin (http://www.webmin.com/) its a
great tool for standard administration tasks
refer to some of these links, they might be useful to get you started
http://www.examcram2.com/articles/article.asp?p=31540&seqNum=4
Some machine have RSC (Remote System Control) capability and others
have ALOM (Advanced Lights Out Management). These provides you the
ability to connect to an internal controller on these systems and do
various remote management type things. Assuming you activate this, you
need to switch your console onto the RSC or ALOM controller. For
systems without RSC or ALOM, you're stuck with serial port console
access which means you need something like a term server at each
location that you can telnet/ssh into.
By creating a matrix of chassie types and their remote managability
capabilities (RSC, ALOM, serial console), you can determine what you
need at each location and start designing your remote console
management network.
Since you are using Sun Cluster, remote console access is critical.
Nothing like Sun Cluster going into a panic loop on a remote system -
such fun.
If using serial port console access, be *SURE* to change
/etc/default/kbd and enable the alternate break sequence. The
CNTRL-BREAK sequence (which is hardware sequence) is the default and
will still be active for a short period during the boot sequence. Once
the kbd file is read, this change to ~<CTRL>B (which is a software
break sequence).
2) You should look into WAN-booting on Solaris as well if you need to
jumpstart a server remotely. This of course requires the design and
implementation of #1 above, OBP updated to support WAN-booting
(assuming the chassie class will support it), and a jumpstart server
somewhere. This isn't really related to patch management but it is
indirectly related in case of a system failure (say from a patch
install - Oh, that's never happened before) and you need to do a
complete reinstall but don't really feel like spending $$ to fly out
and do it in person. This would obviously require some level of
network capacity to allow the installation to proceed at a rate that it
finishes in this century. Oh course, #3 before may also resolve this
little annoyance.
3) Read up on Live Upgrade capabilities - but I'm not sure if LU is
compatbile with Sun Cluster at the moment though. LU allows you to
patch an image of the existing running OS. When complete, you switch
over to the new image and reboot. If there are problems, you simply
switch back to the original OS image which doesn't have the new
patches. Being remote, if a patch installation fails the system
somehow and without remote console access - you're screwed.
HTH
- Next message: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Previous message: mario: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- In reply to: mario: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Next in thread: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Reply: oracle.trainer_at_gmail.com: "Re: Remote administration of a 2-node sun cluster running Solaris 5.8"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|