Transform a Raspberry Pi 3b into an inexpensive 5250 terminal

The Raspberry Pi 3b kit, from various kit providers, costs around $50 including case, heat sinks and power supply. Add a 8GB microSDHC card for less than $10 and you've got the makings of an inexpensive (sub $60) 5250 terminal.

There are two basic methods we've used with good success to create a 5250 terminal from a Raspberry Pi. First, if you have an IBM-i with active SWMA and are running IBM-i 7.x, then the new IBM-i Access Client Solution is one option. Alternatively, if you're wanting a simple user interface with less user options (ie 5250-only) then the open source TN5250j project is your other option. Both of these programs are 100% java-based and thus can run on any OS or device that supports java without having to be recompiled.

The first basic assumption for our initial Raspberry Pi 5250 project is that we do not want a full GUI with all the bloat-ware that comes with Raspbian OS (as of the date of this post the current version is Raspbian Jessie). Thus, we're building this with Raspbian Jessie Lite and adding a basic GUI. The second assumption is that we only want 5250 functionality and none of the user data connectivity offered by the IBM Access Client Solutions application, thus we will use the TN5250J product for this post.

Here are the basic steps that I will outline in some detail in this article:

1. Download software and prep the microSD card
2. Install the Raspbian Jessie Lite OS
3. Update and modify the Raspbian OS
4. Copy 5250 program and modify files
5. Modify the boot load sequence of OS to auto-launch 5250

Step1: Download the Raspbian Jessie Lite .img file zip archive and unzip. If you're a Windows user you'll need to install Win32DiskImager. Also, go ahead and download TN5250J from sourceforge. If you purchase a new microSCHC card it should already be formatted, if not format as FAT32.

Step 2: Use Win32DiskImager to burn the Raspian Jessie Lite .img file to the microSD card. My suggestion is that if you're going to use this microSD as a template for creating more microSD cards for more terminals, use a 4GB microSD card as your template. That way the image you'll create will be 4GB instead of 8, 16 or 32GB. Once Win32DiskImager finishes writing the .img to the microSD card, Windows will prompt you to Format the drive - CANCEL this prompt. Accept and Exit Win32DiskImager. Eject the microSD card.

Step 3: Time to hone your Linux chops or learn some new ones. Once you insert the microSD card into your pi and apply power, the OS should boot to a login prompt. UserID: pi and Password: raspberry.

sudo raspi-config ->(4)localization -> (I1)Locales->Change Locale; scroll down and un-select (using space bar) en_GB.UTF-8 (unless of course you live in Great Britain), scroll down further to en_US.UTF-8 (for those in the US), tab, OK, on the next screen arrow down to en_US.UTF-8 (or other selected locale), tab, OK (this will generate locale files); Next, set your timezone: (I2)timezone (in my case US->Central); localization -> (I3)keyboard (Generic US/104 or other depending on location), next; the system defaults to English(GB), so scroll down to Other, select appropriate option, in my case English (US), next, scroll to top of list, US (English) or other language as appropriate; OK. Next, you should change the default pi password for security; finish (should prompt to reboot); if not, type sudo reboot at the terminal prompt.

Now for installing a base GUI without all the stock bloat-ware:

To make the RaspPi automatically boot to the GUI, use sudo raspi-config to go to Boot Option menu ->(B1)Desktop/CLI -> (B4)Desktop Autologin; tab; OK; tab; Finish;

Step 4: Next, we'll download tn5250j-0.7.6-full-bin.zip (or latest version) from SourceForge and unzip. Rename the unzipped folder to tn5250j for simplicity. Copy the folder to a USB stick. Insert the USB stick into the booted RaspberryPi. The GUI should automatically prompt to "Open with File Manager". Use File Manager to drag the tn5250j folder to /home/pi/ folder.

To create shell script for running TN5250J program:

To enable TN5250J to auto-start when pi boots:

To add an icon to the Desktop for TN5250J, go to a terminal session:

At this point, if you reboot the RasPi the system should auto-login to the GUI and then launch the TN5250J program (if you elected for auto-launch option above). You can then set up a terminal session to your IBM-i system (yes, you can define multiple sessions), remap your keyboard (by default the enter key on the number pad isn't mapped as Enter). You're restricted to two key maps per function, so if you're used to the right control key also being an Enter key, you'll have to choose. If for some reason you should have an issue with the TN5250J application locking up on launch due to a possible error in session settings, just launch a terminal session, cd /home/pi/.tn5250j *important: put the period in front of tn5250j in this case, as it is a hidden directory that contains the configuration files*. Typing pwd will display your current directory. Use ls -l to display files in the directory. You can delete the sessions file using sudo rm sessions. TN5250J will recreate the sessions file when you launch TN5250J and re-add the sessions to the configuration screen.  Always remember to click APPLY.  Mark the first session you create as "default" and it will auto-launch that session every time the device boots.  Also, to modify session or add sessions, you must right-click the mouse inside the session window and the edit options panel will pop up.

If you want the Display Power Management System (DPMS) to NOT auto-sleep the display (ie the screen stay ON all the time), add the following line to the lightdm.conf file:

cd /etc/lightdm
ls -l
(should be a file called lightdm.conf in that directory)
sudo nano /etc/lightdm/lightdm.conf
arrow down to the [SeatDefaults] section of the file
add the following line underneath
[SeatDefaults]
xserver-command=X -s 0 -dpms

^O, enter to save, ^X to exit

Another "nice to have" feature would be to enable VNC via the raspi-config function so that you can provide remote support for the RasPi terminal. There is also a Citrix Receiver built for the Raspberry Pi as well as a Parallels 2X client, RDP client, and several browsers including Chrome and Firefox. Setup and installation of those are a subject for another post.  You can also set the top Panel bar to AutoHide so that if you're using the device without a mouse the panel bar doesn't show at the top of the screen.

In case you were interested, my typical hardware configuration as of this post consists of a RaspberryPi 3b board ($34.99), an Enokay case ($5.99) which screws together and comes with the two heat sinks needed for the pi board, a NorthPada power supply with on/off switch for $8.99 and an 8GB SANDisk microSD card which in bulk costs around $5.50.  Don't forget an HDMI to VGA or HDMI to DVI converter cable if your monitor(s) do not have HDMI input ports.

Brocade Fiber Channel Switch administration and Java support in new browsers

Older model Brocade Fiber Channel switches including IBM OEM Brocade switches use a Java webstart applet (Java Network Launching Protocol or JNLP file) for administration.  If you've tried to administer a Brocade Fiber Channel switch lately with a newer browser, it is likely you've been denied access.  Chrome, Firefox and IE have all discontinued support for NPAPI plugins, including Java webstart.

Pale Moon to the rescue!

The Pale Moon browser (from palemoon.org) is a fork of the Mozilla/Firefox code and has not (as of the date of this posting) removed the NPAPI Java plugin support.  While they do recommend that you DISABLE this feature for security purposes, the ability to associate a JNLP file with a java webstart executable and run java applications in the browser via the NPAPI plugin is still supported.

To resolve your problem with administering your Brocade Fiber Channel switch using the Java applet, do the following (note: this assumes you already have Java installed):

1. Download and install the Pale Moon browser
2. Reboot your machine
3. Launch the Pale Moon browser
4. Point the URL to the IP or DNS address of your Brocade switch
5. When prompted on how to launch the JNLP file, browse to your Program Files (or Program Files x86 if 32-bit) Java directory, select the java version you want to use (directory), go to the bin directory, and select the javaws.exe as the executable for the JNLP file.  For example: C:\>Program Files\Java\jre1.8.0_91\bin\
6. If you receive a java security error, go to the Java app properties, go to the Security tab, and add the DNS or IP entry to the Exception Site list.  Don't forget to put http:// or https:// in front of the entry.

I've had this solve my Brocade switch java access problem on IBM 2498 series switches including the 2498-B24 and 2498-24E (Brocade 300-series), as well as the Brocade 5470-based BNT Fiber Channel 10 and 20-port switches for the IBM Bladecenter Chassis.  This also works on the Cybernetics VTL administration control panel for Cybernetics Virtual Tape Library devices.  I have had the case on one laptop where I had to go into the file associations (I think it was a Windows 8.1 system) and associate the JNLP file type with the javaws.exe for execution. 

Move Your File Server to the "Cloud"

• This is a re-post of an article I wrote on LinkedIn:

  I'm asked frequently about the costs and process of moving a small office file and application server to the cloud. At Akzium we have moved entire datacenters with hundreds of servers to a cloud-hosted environment and we've helped small businesses avoid the upfront investment costs of a physical server by leveraging low cost cloud computing options.
  In this post I'll outline a very simple way that a small office can "go cloud" with their file and application server. This is just one way to accomplish the task but it serves as a nice blueprint for all of the other variations. Also, while there are many options out there for cloud hosting, I'm going to use Amazon Web Services (AWS) here as the example since their pricing is easily accessible.
  
  Here is a brief outline of the steps involved in this process:
  

  • Sign up for AWS and create an EC2 instance with EBS storage (~$65.00/mo)
  • Configure the AWS EC2 Windows Server (updates, installed antivirus, opened ports 80/443/3389 on the EC2 instance firewall and VPC
  • Create folder(s) on the AWS EC2 Windows file server and assign share permissions.
  • Sign up for LogMeIn Hamachi account ($29.00/yr)
  • Install Hamachi on the AWS EC2 Server and designated it a HUB
  • Install Hamachi on each PC/Mac needing access to the server and designated each of them as a SPOKE
  • Used the "browse" option in the Hamachi client on the PC/Mac to connect to the HUB and login using Windows server username/password to gain access to the permissions.

  
  With those steps we've created a cloud-hosted Windows file server with VPN access from remote PC/Mac clients for file shares. I would also suggest installing something like CloudBerryLab's cloud backup for servers ($79.99 one-time-cost) and doing backups of all shared folders into Amazon S3, Azure or any of the many supported cloud object storage providers supported by CloudberryLab. Costs vary by provider but AWS S3 is roughly three cents ($.03) per GB per month, so for our 120GB example that would be $3.60/month in backup costs if the entire 120GB was used.
  
  For those who want a more detailed description, I'll expand upon the steps listed above. First, we need to set up a cloud server. In this case I'm going to use a server with Microsoft's Windows Server 2012r2 since most businesses use Windows servers, but for the tech savvy user AWS also offers Linux which can lower one's monthly hosting costs. To get an estimate of costs for this post I'm using Amazon's AWS Simple Monthly Calculator for each of the AWS items we'll be discussing here. Also, I'm not intending for this to be a step-by-step whitepaper on the process but rather to outline the general procedures for a cloud hosted file server. An AWS EC2 instance labeled "t2.medium" provides 2 cpus of compute power and 4GB of memory for $52.71/month including the Windows Server 2012r2 operating system, plenty of juice for a small office file server. EBS storage for the file server is roughly five cents ($.05) per GB per month. For this example let's assume a need for an 80GB "C:" drive and a 120GB "E:" drive for a total of 200GB of EBS storage which will cost and additional $10.00/month. Make sure to assign an AWS Elastic IP to the EC2 instance so that you have a static public IP address. There are additional bandwidth charges so I've estimated 200GB/mo. of total data transfer which costs and additional $2.00/month for a grand total of $64.71/month. In the interest of keeping this example very simple I'm not going to use Amazon's VPC firewall to set up a site-to-site IPSEC VPN tunnel but rather am going to opt for LogMeIn's Hamachi VPN solution. Only very limited technical knowledge is needed for this type of VPN whereas more technical networking skill is required to set up the IPSEC VPN. The t2.medium instance type does require that you configure the AWS VPC and assign the instance; you'll also need to open ports 80, 443 and 3389 on the VPC security rules. Again, I'll not go into a deep dive tutorial here on AWS EC2 instance creation and VPC setup here as there are plenty of resources to assist with that elsewhere.
  
  LogMeIn's Hamachi is a simple solution for creating a virtual network. You'll need to create a LogMeIn account and log into the control panel. For $29.00 per year you can set up a single site VPN with up to 32 remote users. The LogMeIn Hamachi client is installed on each PC or Mac and from the LogMeIn control panel you designate network type, members and settings. There are some step by step guides out there and this post is not intended to guide you through the setup but rather as an introduction to a reference architecture. Setup of the Hamachi network is rather simple, once you login to your LogMeIn account, click on the left hand option for "My Networks", select "Add Network", give your new network a name such as "XYZ Co Network" and a description and in this example I elect to use the hub-and-spoke option where the cloud-hosted server is the "hub" and all other devices are "spokes." Now, go login to the AWS-hosted server (RDP), open the browser, go to LogMeIn, login to your account, choose "My Networks" on the left and install the Hamachi client via "Add Client." The "trick" to connecting to your recently created network is to know the network ID number which you can get by clicking on the Edit button next to your recently created network. The ID will be a 9-digit number (eg 123-456-789). I elected in my Network Settings to require a password and approval before a member can be added to the network so if you elect that security option you'll have to go to the LogMeIn control panel and approve the member. Now, the one rather non-transparent operation in the Hamachi setup is the process for designating a "hub" in their hub-and-spoke network configuration. To do this you go to Add/Remove members and the Hub and Spoke radio buttons appear on this screen allowing you to designate your recently approved network member as the Hub server. Not at all intuitive but once you understand where it is it's not an issue. Next you should go install the Hamachi client on each Windows PC or Mac that you want to make a member of the network, again approving the addition of each member via your control panel.
  
  Once you've set up Hamachi on the AWS server and designated it a Hub, you'll need to create folders and share them just as you would on a locally created network server. If you elect not to install the Domain Controller role on the AWS server you can simply create a "workgroup" type network. If you go the workgroup route, to make things easier you might want to change the workgroup name on all of the PCs to match your designated workgroup name on your AWS server as well as changing the AWS server instance name to something besides the default cryptic WIN-xxxxxxxxxx naming convention. If you are restricting shares by user permissions you can go to the control panel, add users to the local machine (in this case our AWS server) and designate folder permissions based on user. I won't delve into a Windows user permissions tutorial here but suffice it to say that you should do everything just as if you were on a local server. Next, go to one of the local PCs that you've installed the Hamachi client on (a "spoke" in Hamachi-speak) and right click on the Hamachi icon in the Notifications area, open the Hamachi client and you'll see your Hamachi Network, you should now be able to click on the network and "see" the AWS server listed. Occasionally, I've had to reboot the client before it recognized the network. If you click on the "browse" option, a local Windows File Manager will open and you should be prompted to login to the server, just remember to use the XYZSERVERNAME\username convention in the UserID portion of the login screen if you're not on a domain. Once you've authenticated to the server you should see your shared folder(s) listed. You can then map a drive letter to the folder(s). If you make the AWS server a Domain Controller things get a little easier but that can also bring up other issues we won't make a part of this post.
  
  A twist on this solution would be to keep a local file server and use the EC2 instance as a cloud hosted replica using Microsoft's DFS replication solution or a third-party option such as Vision's DoubleTake or even the CDP option in CloudberryLab's backup software. In the event of a local disaster you could then instantly launch the Hamachi client and connect to the replicated EC2 file server and continue on with business as usual until the local server was repaired or replaced.