I’m about to move some web servers from a dedicated CentOS LAMP box at a traditional hosting environment to Amazon Web Services.

There is quite alot to consider and there seems to be more architecture planning involved in running a system there at Amazon’s service and a little bit different way of doing things than the traditional web hosting environment, but it also opens up a great potential to testing out dev servers or rolling dev servers into productions servers, etc. etc. not to mention the scalability and redundancy offered.

While doing research, I keep seeing more and more and more about Ubuntu, which sparked me to do a little research and see what all the hubabub was all about.

I’m now considering running Linux Ubuntu 10.04 Lynx server as my server instead of CentOS.

Does anyone out there in the Lotus blogosphere have any experience with moving from Fedora/CentOS to Debian/Ubuntu flavors? If so, I’d love to have a casual chat. More importantly, is anyone using Amazon Web Services? It seems like quite a considerably more complex operation than throwing a few physical boxes in a rack and being done with them.

I just recently signed up for LinkedIn and joined a few Notes/Domino groups. I haven’t done much with it yet, but here’s my public profile http://www.linkedin.com/in/davidkillingsworth.

Today, I received a LinkedIn post with an answer which pointed to Technote 1196837 – Using a Desktop Policy to set notes.ini and Location parameters.

I really wish I had seen this technote a very long time ago.

I’ve been playing with the HP Systems Insight Management Console for the last few months. It’s a great tool and lots and lots of granular monitoring, patch and firmware deployment, etc.

I just realized this week that you could go to HP’s website and install what’s called a Proliant Support Pack.

make sure and set some settings in SNMP or the agents won’t work when you load up HP Systems Management Homepage.

You’ll need to go to the Service Panel and right click on SNMP Service. On the Traps Tab, you’ll need to specify a community of public (or whatever might be the standard at your organization).

Go to the security tab, and check the “send authentication trap”
For public rights should be “READ ONLY”
for private rights should be “READ CREATE”

In the “Accept SNMP Packets from these hosts” you will need to specify 127.0.0.1 and localhost.
If you have a HP Systems Insight Management Console setup, you’ll need to specify the hostname or the IP address of your HP Systems Insight Management Console server.

These settings may be set for you if you use the HP Smart Start DVD to install Windows 2003.

If you need to install SNMP after you install Windows, be careful as it automatically selects IIS server when you select SNMP when adding programs to get the SNMP service installed.

I also had a bit of trouble installing SNMP after Windows 2K3 Service Pack 2 had been installed. I had to insert a Windows 2003 Server CD, and also download and extract the SP2 install image to a directory because the SNMP installation required files from both install packages before it would install.

I tricked SP2 by clicking on the download image and letting it expand, and then copying the expanded folder to another location and then clicking cancel, which then deleted the original expanded SP2 install directory.

I have since found out that there is a command to expand it.
WindowsServer2003-KB914961-SP2-x86-ENU.exe -x:c:\temp\sp2\
to expand it into the c:\temp\sp2\ directory.

Now that you have SNMP and HP Systems Insight Management Home Page working…you can check to see if your hardware has any problems, what type of HP system you have and what and how many of each component you have.

If you have a HP Systems Insight Management Console, Version control agents (on the client server), and an up-to-date Version Control Database (on the Management Console server), you can determine what version of ROM and drivers you have compared with what has been downloaded from HP on your Insight Management Console and is available for install from your Insight Management Console server.

This is a great way to keep your server up-to-date.

Additionally, you can “push” the latest updates to a single or list of (client) servers from the HP Insight Management Console servers. It works very well, and its very simple once the whole thing is already setup. Sometimes discovery can be a little flaky and you need to specify your servers on the HP Systems Insight Management Console, but otherwise, it’s very good.

But what if you don’t have a spare server laying around to use as a HP Systems Insight Management Console, with the Version Control Database on it?

You can download the Prolian Support Pack as mentioned above for your server and apply all the latest baseline HP software and Windows drivers to a single server or even push that to a group of local servers.

It’s kind of the same concept without the dedicated HP Systems Insight Management Console server.

One thing to note is that after you download and install the Proliant Support Pack from HP, you’ll also need to download the ROM updates for the CPU, ILO and ILO2 Boards, and Array Controllers seperately from the Proliant Support Pack. Once you have downloaded the correct ones, you can put them in the same directory with the other Proliant Support Pack files and apply them all at once.

Very cool stuff.

One thing that I have to complain about though is that it is VERY VERY difficult to find individual downloads on the HP site. It’s almost impossible even using google and their own search engine.

Here’s a very useful URL. START HERE -> http://h20180.www2.hp.com/apps/Nav?h_pagetype=s-001&h_lang=en&h_cc=us&h_product=241435&h_client=S-A-R163-1&h_page=hpcom&lang=en&cc=us

I haven’t.

I sat down to register 284 users this week. It was a new process to me, so I’m going to go through the list of steps here.

1. Created a new OU – let’s call it WHEAT.
2. Edited the Recovery information for this OU ID.
3. Created a group for all users on this new server cluster.
4. Organized a list of users given by HR to be registered in excel spreadsheet format.
5. Organized the columns according to the Domino help files.

I used the following columns, though there are more you can specify in the help file.
lastname,firstname,middle,,,id file directory,id file name,server,directory,mailfile,location,,,,local administrator,internet address,shortname,,,mail template

6. Exported the Excel file to a .csv to get the comma formatted text above.
7. Edited the .csv file with a text editor, removed the header column line (line 1) and replaced all commas (,) with semi colons (;) as required by the import file format.

8. I went to the server where I was going to register all users, created a replica of the certification log (certlog.nsf) and our Recovery Database (mail\recovery.nsf).

9. I set my administration preferences to use the server where I would register the users, Configuration\Tools\Register\Person.

10. I set the Registration Server, chose advanced, set the Internet address field, set the certificate expiration date to more than 2 years.

NOTE: This step is important, because all of the imported users from the list will inherit anything you set here.

11. Imported the text file.

12. Selected all users in the imported list and chose to create a replica of their mailfile on the cluster mate server in the cluster.

13. Selected all users in the imported list and deselected Create ID file in address book, and selected “create ID file in “location” where the location was already specified in my import file.

I didn’t have to set anything else. I was able to make the short name, ID file, and first part of the Internet address all the same by specifying each in the import text file.

There were some things that could be set in the import file, some things that needed to be specified after import, and some things that probably could have been set through a registration policy. I didn’t see any real value in setting up a registration policy this time.

All said and done the generation of ids, person documents, and mailboxes is still in progress, but I’m guessing it will last about 2 hours.

Who uses anything besides TCPIP anymore?

I always disable any non TCPIP ports after install as it complicates Notes configuration for desktop support people, complicates location document creation, and I had one user once who had all the default ports enabled. He also had some network drives permanently mapped. When he was at a site that could not see the network mapped drives, and he would save an attachment, Notes would freeze.

After some study, I realized that Notes was trying to build the list of places to store the attachment for explorer and searching every possible port to map those drives.

I disabled everything except TCPIP for the guy and tried the same process. The guy thought I was Santa Claus and that he’d just received the best Christmas gift ever. The attachment save dialog window instantly opened and presented a local directory.

I’ve tried several ways to get around this on install (and there is no way to disable the ports by policies).

1. The installshield tuner has a bug where you can’t disable the “Optional Network Drivers” Feature. You can only select, “The feature is installed on the local drive,” and “The feature is run from source, CD, or Network.”

2. I’ve also tried setting these in the notes.ini in the InstallShield Tuner package.
Ports=TCPIP
DisabledPorts=LAN0,COM1,COM4,SPX,COM2,COM3,COM5

After installation finishes, those entries are in the notes.ini. However, as soon as you launch Notes for the first time and run the setup, the setup program overwrites those notes.ini variables and adds Lan0 and Com1 and/or Com3.!!!!

3. I’ve also tried setting configfile=c:\temp\setup.txt in the notes.ini on the InstallShield tuner, so that this line is added to notes.ini on installation.

Then during installation, Notes looks at c:\temp\setup.txt
In setup.txt, there are a few variables you can set. One of them being Domino.Port. I thought by setting Domino.Port=TCPIP, it would only install TCPIP, ….NOT..

I finally figured out how to disable the installation of Optional Network Drivers in InstallShield. It seems like it’s a bug in the InstallShield UI.

If you choose Additional Tools\Direct Editor\ then Feature, then Optional Network Drivers. Set the “Level i2″ field to “101″ and the “Attributes -i2″ to 9.

You can go back up to the Organization\Features UI section for Optional Network Drivers, and you will now see that the initial state is now set as “The feature is not installed.”

Why was this so hard to disable ???