Linux for single-purpose terminals

Here's a little Arch Linux success story that I feel I need to post somewhere:
I work at a place that oversees public internet access sites (government-funded, for disadvantaged people but anyone can use them), and we're often pressured into making the terminals run Windows "because that's what people are used to." I've about had it with trying to get Windows locked down enough so that people can't set porn wallpaper or change the homepage, but can still save PDF's to the desktop and print them. Securing windows is an adventure in clicking through hundreds of little checkboxes -- it starts with everything allowed, and you selectively turn things off one by one, and hope that Microsoft deigned to make checkboxes for everything.
And then, there's the problem of cloning the install once you've got it all set up just right -- even on identical hardware, you can't just copy the files over and expect it to work. Even after you run the repair CD and use that "fix boot sector" command, usually it still won't boot without a little more fighting, and most of the time a "repair install" is needed. When you've got different motherboards, especially if they're of a different generations (686 vs. one higher), almost nothing is portable.
When I'm not making Windows boxes for these sites, I have sysadmin duties that I often don't work enough hours per week to fulfill. And when an order for "make us some public terminals" comes in, they usually mean "right freaking now" and we don't have time to do anything but fulfill the request the only way we know how, so I don't get time to experiment then, either.
Then the other week, one of my contract clients from my other job (my own business, actually) said, "We need four very simple terminals that act as web browsers for this web-based app our volunteers will use. We want them as cheap as possible, so it'd be great if you could use junk hardware."
So, I finally had a legit excuse to do what I'd always wanted to -- build these things with Linux. I chose Arch because that's what I use at home, so I'm already familiar with it. It's also insanely fast to get it installed and running -- if you use the "quickinst" method, you don't even have to bother with package selection until later. I built the first workstation, then spent a few hours Googling for how to make X auto-login as the limit user, then how to make a custom UI out of FVWM so that all the users had were gigantic buttons called "WWW" "My documents" "Notepad" and "Citrix" and "Help". I found an XP-like theme for FVWM, so that users who've memorized "big red button means close window" will be happy, and learned enough FVWM scripting to have it customized within a couple hours. I found that opera-static (web browser) loads fast enough, even on a Celeron 466 and a really old hard drive.
Then came cloning -- I weeded out as many unnecessary files as I could, and made the root partition only use 780 MB uncompressed. I made tarballs of it and /boot , booted the next blank computer from CD (totally different hardware -- my initial machine was a P4, my clone was that Celeron 466; different hard drives, network cards, video, everything), partitioned it, untarred, changed the hostname and 2 or 3 other little settings, installed grub... and the thing booted perfectly on the second try (the first try was spent switching X drivers). I repeated the process to the other two machines, and had them all working by the end of the day. Oh, and the Linux version of the Citrix client is faster than the native Windows one :)
Four machines with vastly different hardware, running identical setups that (I hope) will be damn near impossible for users to screw up, in one working day. It took me and the staff at my day job months to learn all of the tricks for doing the same thing with Windows, with less adequate results. Now I'll be sharing my linux station tarballs with them, and we can make workstations in less than 10 minutes, rather than 3 or 4 hours, and best of all, no more getting into arguments about licensing (it's become more or less standard practice to use pirated Windows even though we have paid keys for every install, just because it saves us so much hassle -- no WGA, no activation, etc.).

thetrivialstuff wrote:no more getting into arguments about licensing (it's become more or less standard practice to use pirated Windows even though we have paid keys for every install, just because it saves us so much hassle -- no WGA, no activation, etc.).
Don't you love it when they shoot themselves in the foot so readily?
Very nice post

Similar Messages

  • Is it possible to download current Red Hat Linux for learning purpose?

    Hi,
    Is it possible to download current Red Hat Linux for learning purpose? On their website, you either have to purchase the regular version or download a 30-day evaluation version.
    Thanks.
    -c

    How different is Enterprise Linux compared to Fedora?To add Sergio's post, Fedora and Enterprise Linux are based on RedHat distribution. Fedora is non commercial distribution and is very popular. Enterprise Linux is quite new (name of distribution) in the Linux world (announced at the end of october) and it is clone (as Sergio said) of RH EL 4 U4.
    All I need is an opportunity to learn Unix/Linux.Any of Linux distributions will give to you this opportunity. And for learning of Linux Fedora is good enough.
    Is Enterprise Linux more popular?
    Popular than what? Fedora? It is too soon... (IMHO). Enterprise Linux is ideal candidate for productions systems which needs official support and (of course) systems which where Oracle products will be running on. In this scope (commmercial sphere) Enterprise Linux will become popular than Fedora (which has support only from volunteers) but I don't think that Enterprise Linux will be more popular as Fedora at all...

  • Multiple ud code appearing for single purpose?

    Hi all,
    When i am doing usage decision for goods receipt, multiple UD code are appearing, with any code also UD is happening - for example - with A 01 also and YB01 01 also, can you please suggest me is there any different purpose for these codes and how do i enable a single code.
    regards,
    Sanju

    if u want some specific ud codes should appear while taking ud
    create selected in set in qs51 for that plant.
    as said in above thread go to  Quality Management > Quality Inspection > Inspection Lot Creation > Maintain Inspection Types > Detail of each inspection type -in selected set enter this newly created selected set and set indicator 'Selected set in same plant'.
    only those codes will appear now
    Regards

  • HT204053 I had Snow Leopard and was using iWeb and Filezilla for my website (not MobileMe). Wanting to move to single-click publishing, I now find it is not supported by iCloud and Mountain Lion. I feel cheated, having bought it mainly for this purpose!

    I had Snow Leopard and was using iWeb and Filezilla for my website (not MobileMe). Wanting to move to single-click publishing (supported by MobileMe), I now find it is not supported by iCloud and Mountain Lion. I feel cheated, having bought it mainly for this purpose! The other thing they don't tell you is that Mountain Lion disables OfficeMac, and I am considering uninstalling it for that reason - do I get my money back?!

    You stated; "Mountain Lion disables OfficeMac"
    That is not true.
    Mountain Lion does not have Rosetta so it is not capable of executing PowerPC code. If you have MS Office 2004 that is coded in PowerPC code and will not run in Mountain Lion. What you need to do is upgrade to an Intel version of Office.
    Allan

  • Install a grid infrastructure for a cluster for single node Windows Server

    Hello,
    Can you suggest me how to install a Oracle Single Node RAC 11g on a Windows Server 2008 x64 bits? It's for testing purposes and I don't need two nodes. I know that Linux is better, but my company only uses Windows servers. Our former consultant was able to do it, but didn't tell us how to and that server was lost.
    I know that we have to install first Grid Infrastructure. Running the "Oracle Grid Infrastructure" installer, I could successfully install a "Grid Infrastructure for a Standalone Server". But that grid is not for a RAC database. When trying to install the Database software, it allowed to me to only install "Single instance database", not "Real Application Cluster database installation".
    Running the "Oracle Grid Infrastructure" installer, I tried to select option:
    "Install and Configure Grid Infrastructure for a Cluster" -> Typical Installation:
    SCAN Name: SRVORAC-cluster
    Validating SCAN information
    INS-40922 Invalid SCAN Name - unresolvable to IP address.
    Is it possible to install a grid infrastructure for a cluster for a single-node or one-node Windows Server ?
    Edited by: user521219 on Jan 30, 2012 12:46 PM

    HI,
    You probably have misconception about Single node RAC. This feature does not need that you have only one physical machine in a RAC, it actually means that in a RAC environment (2 or more nodes), you can run your database on only one node(single instance) and then later you can move this single instance to any other node of the RAC (so that you can do some patching on this node making your database available all the time). This also has some other benefits, see bellow.
    http://www.oracle.com/technetwork/database/clustering/overview/ug-raconenode-2009-130760.pdf
    http://docs.oracle.com/cd/E11882_01/install.112/e25666/whatsnew.htm#sthref8
    Salman

  • Exchange Server 2013 SP1 - Optimize for Single Server Use

    Hello,
    i've set up Exchange 2013 SP1 with at last CU6 - get exchangeserver says:
    AdminDisplayVersion             : Version 15.0 (Build 995.29)
    ExchangeVersion                 : 0.1 (8.0.535.0)              - - - funny number ;-) (is this an early version???)
    Yes, the annoying 16028 in the application log are running every five minutes, but thats not the main question.
    Which services are needed in a SINGLE-SERVER Environment, without DAG and unified messaging?
    Can the Self-Checking, Health- and Self-Probing be put down?
    These services are started automatic:
       Microsoft Exchange Active Directory Topology
       Microsoft Exchange Anti-spam Update
       Microsoft Exchange DAG Management
       Microsoft Exchange Diagnostics
       Microsoft Exchange EdgeSync
       Microsoft Exchange Frontend Transport
       Microsoft Exchange Health Manager
       Microsoft Exchange Mailbox Assistants
       Microsoft Exchange Mailbox Replication
       Microsoft Exchange Mailbox Transport Delivery
       Microsoft Exchange Mailbox Transport Submission
       Microsoft Exchange Migration Workflow
       Microsoft Exchange Replication
       Microsoft Exchange RPC Client Access
       Microsoft Exchange Search
       Microsoft Exchange Search Host Controller
       Microsoft Exchange Service Host
       Microsoft Exchange Throttling
       Microsoft Exchange Transport
       Microsoft Exchange Transport Log Search
       Microsoft Exchange Unified Messaging
       Microsoft Exchange Unified Messaging Call Router
       Microsoft Exchange-Informationsspeicher
       Microsoft Filtering Management Service
       Microsoft Online Services Sign-in Assistant
    And with C:\Program Files\Microsoft\Exchange Server\V15\TransportRoles\data\Queue\mail.que about 512 MBytes size (which may may be a designed default size)...
    Have set up about 6 users mailboxes, 3 public folder in a mailbox for public folders. Only for testing purposes sent some (small) mails, no external mails received (only testing receive with 'popcon' for mail delivery to the internal boxes).
    The server (2012 R2) is running only in the internal network (yes, with internet connection for updates).
    No virus scanners, not third party receive or send connectors, only the 5 default (3 Frontend, 2 Hub) receive Connectors and one Send Connector to our mail provider set up and no user activity (only sometimes from me for testing).
    It seems like it's doing only with itself, logging file sizes of above 11 GB (in 4 weeks), with huge amounts of probing, health checking, internal testing, internal updating etc.
    Can this logging be set up to a lower level, can services be deactivated (for single server configuration),
    the logging volume should be in a useful relation to the mail use of the server, a message for every self test, internal health check etc. in the log-files normally isn't useful, only when problems occur there is a need...
    Thanks in advance
    Andreas

    Between Exchange protocol-based log files, PerfMon .blg files used by Managed Availability, IIS logs, etc., the total footprint can add up quickly and these logs are not purged automatically.  You can take a look at this site below and use the powershell
    script to clean it up.
    http://www.c7solutions.com/2013/04/removing-old-exchange-2013-log-files-html

  • Grouping of vendor's for reporting purpose

    Dear Experts,
    i have an issue where i have different vendor code (For example - IBM- Mumbai, IBM-Bangalore,IBM- US,IBM-Germany). i have created different vendor code for all these companies.
    Now i want to link all these vendors under single grouping code  so that i can see the report based on group company "IBM" that what all the business i have done against the group company "IBM".
    Is there any concept of grouping several vendor's under single code for reporting purpose.
    Plz suggest.

    The best way to resolve your issue i think is by creating a vendor class with class type 010 in t/code CL02...before this create a characteristic where maintain the Vendor codes & name in Values data tab in t/code CT04 and assign the same in CL02...now you can use this class in any vendor related report by using the option Vendor by class.
    Regards,
    Indranil

  • Correct Hardware Requirment for Educational Purpose

    Goal:
    Buy a new hardware for SharePoint development.
    Purpose:
    Increase my knowledgein SharePoint development. Today. I'm a junior SharePoint developer. I have experience in dot.net deveopment but not in SharePoint development.
    Problem:
    1.
    Is Hardware requirements—database requirement below enough?
    2.
    Which software requirements I should use?
     - Single server with built-in database,
     - Server farm with a single server in the farm or
     - Server farm with multiple servers in the farm
    for a educational purpose. Each of these has different hardware requirement. (http://technet.microsoft.com/en-us/library/cc262485.aspx)
    Information:
    I have consider buying a new desktop och laptop computer with following criteria.
    laptop with hardware requirement:
    16 GB RAM
    Intel processor i7
    Desktop with hardware:
    24 GB RAM
    Intel processor i7
    I gonna install windows 2012, SQL server 2012, SharePoint 2013 in VM Player's environment.

    For development purposes, if you leave out SharePoint Search, you can get away with as little as 10GB RAM allocated to the SharePoint + SQL VM (given you restrict SQL to 1 or 2GB memory).
    If you need Search functional, then go no lower than 16GB for a SharePoint + SQL VM.
    Also make sure you get a separate disk for the SharePoint + SQL VM, SSD if you can swing it, otherwise, in a laptop, a 7200 RPM drive minimum.
    Don't forget that you also need a Domain Controller in your environment running Active Directory. This should not be the same VM as the SharePoint server, but rather a separate VM (for many reasons). This VM can be small, with as little as 512MB RAM allocated
    to it.
    Trevor Seward
    Follow or contact me at...
    &nbsp&nbsp
    This post is my own opinion and does not necessarily reflect the opinion or view of Microsoft, its employees, or other MVPs.

  • Get a test credit card for testing purposes?

    Hi everyone,
    this is Robin from Germany. I work for HUGO BOSS and I currently evaluate Apple Pay for our company-owned stores. We are thinking of upgrading our stores with NFC POS Terminals in order to accept Apple Pay.
    Unfortunately, since our departement is located in Germany we were not able to get a credit card which is working with Apple Pay.
    Is there any possibility to get one for testing purposes?
    Would help a lot.
    Best regards,
    Robin

    As Apple Pay is only in the U.S. right now, why not deal with the Hugo Boss folks in the U.S. to do the testing?  I found this on the hugoboss.com site:

  • O365 federation for testing purposes

    hello,
    could i set up federated trust from my on premisse infrastructure to o365 (using adfs and dyrsync and localy issuated certificate) just for testing purposes , and after test period i would just brake the trust. Could that have any circumstances, because
    in a month or two i would like to migrate to o365?
    regards,
    Mario

    You can absolutely set up Exchange Online and treat it as a separate organization, create a federated trust and run through testing.  It is important to remember that "separate organization" means not using your existing SMTP domain, so you'll likely
    want to avoid ADFS/DirSync for the test period in this scenario.
    An alternative to consider would be setting up Office 365 in a hybrid coexistence configuration, requiring ADFS and DirSync, but putting Exchange Online in a scenario where it operates in tandem with your local Exchange organization.  This coexistence
    would allow you to test with your existing, production SMTP domain and a single GAL view, while still allow you to break off permanently without impact to your production Exchange environment if you chose to do so.

  • How to layout beans for single page design

    Hey guys. I'm designing a site that uses the Icefaces framework. I've been reading a book called Real World Java EE patterns. I'm kind of confused how to layout the pages. Normally I would have just a POJO class implement serializable for a bean. This bean would then back each page. With a single page design I'm going to have a bunch of elements on the page. Datatables, trees, inputs, calendars etc. Is it normal or best practice to have separate beans for each datatable, calendar, etc or put that all in one bean? I'm not sure how to approach this. Right now each element is a bean and I'm using the @Inject annotation to have the data table talk to the tree and vise versa. This creates really bad code and if I put this as a member of the class I will get a circular reference because the data table bean has to inject the calendar and the calendar has to inject the data table.
    Thanks for any help.

    'best' practice is to not follow blindly what other people say, but to reason yourself what works best for you. It has been stated many times before and I will state it again: best practice does not exist. Only personal opinions about what is best exist. Personal opinions are not facts.
    So choose what works for you. Is the code easy to maintain? Is it readable? Can you perhaps re-use parts of it? Then I'd say you have a winner.
    I'll give you my personal favorite. When working with a page that has input elements, I tend to have two or three beans. Replace XXX with a logical and context sensitive name.
    XXXForm - this is a simple request scoped bean that has the action and event methods and will hold some data for reading purposes for the view, possibly taken from a database.
    XXXStorage - this is a session scoped bean that holds information from the form; if it is an editing function I initialize it with the current state first. Storing this stuff in a session bean makes it very easy to do Ajax stuff and create input cycles that span multiple requests. Also when data is split in multiple pages, the storage class is the one to remember which page is being displayed.
    Optionally, I also have an XXXSearch object, which is a session scoped bean that stores search/sort/filter parameters. I make this session scoped so the information is remembered for the duration of the user's visit, meaning he can navigate away from the page in question, come back and all his choices will still be in place.
    So no, I don't have beans specifically for each component, although I might make simple POJOs to hold information gathered from several sources (for example, information gathered from several database tables). These beans are generally used to fill the rows of a datatable and are just created from within the form class, for example in a @PostConstruct annotated method.

  • Is Standard Edition free for development purpose

    Hello,
    I am using Oracle 10g express edition on a 32-bit Redhat Enterprise Linux 4. My test database have come upto 2gb and after that the database stopped working. I have been told to change to 64-bit OS through forums because of the 2gb file size limitation for 32-bit OS. When I checked the Oracle downoad site, there is no 32-bit Oracle express edition 10g. There is a 64-bit Oracle standard edition. Is Oracle 10g Standard Edition free for development purposes? What about Oracle 11g Standard edition?
    Edited by: Thunder44 on Aug 25, 2008 1:09 PM

    All Oracle software, though, is available for download from OTN with a license that allows you to prototype your application. Once the application is beyond the prototype phase, you'd need to purchase an appropriate license for your development environment, but you may just be looking at prototyping...
    As well, note that the database size and RAM restrictions in the express edition would not be expected to change if Oracle ever released a 64-bit version. The whole point of those restrictions is to ensure that the free express edition doesn't cannibalize sales of the standard edition databases.
    Justin

  • Can I use Creative Cloud Photography Plan for commercial purpose?

    I own small company and we have some basic use of Photoshop. I am planning to buy creative cloud photography plan. Can I use creative cloud photography plan for commercial purpose?

    By using it for your own profession to be used only by you, you can choose between any option but to get the CC used by your staff or other member of the organisation, Team needs to be used.
    If you are using the retail for your own profession then you get 20GB or 2gb as per the individual plan purchased where as CCT with single seat with provide you access with 100GB of storage space.
    Regards
    Rajshree

  • Merge Flash Player Skin with Flash Video for Single Flash File

    My company uses a content management system for it's website and I am the administrator. The CMS does not require or use html code. I am a new to Flash, but I am able to produce some simple files. I have compiled a simple 2 minute video clip that I need to upload to my companies website. The Flash file requires a player skin which I was able to produce. I understand that when you output to Flash, two files are created. As I understand it, one file is for the skin and the other is for the movie along with the html code that links the two. Unfortunately, our content management system does not recognize the link between these two sets of files and I am unable to import them. I can import the flash movie clip without the player, but that does not suit my needs. I can import the player skin without the movie, but there would be no point. Is there a way to output both files into a single Flash file with the player skin embedded into the Flash movie? I have the full Adobe Master Suite CS4, is it possible another program could be used for that purpose or is this an impossibility? Any information you could provide would be appreciated...thanks...

    My company uses a content management system for it's website and I am the administrator. The CMS does not require or use html code. I am a new to Flash, but I am able to produce some simple files. I have compiled a simple 2 minute video clip that I need to upload to my companies website. The Flash file requires a player skin which I was able to produce. I understand that when you output to Flash, two files are created. As I understand it, one file is for the skin and the other is for the movie along with the html code that links the two. Unfortunately, our content management system does not recognize the link between these two sets of files and I am unable to import them. I can import the flash movie clip without the player, but that does not suit my needs. I can import the player skin without the movie, but there would be no point. Is there a way to output both files into a single Flash file with the player skin embedded into the Flash movie? I have the full Adobe Master Suite CS4, is it possible another program could be used for that purpose or is this an impossibility? Any information you could provide would be appreciated...thanks...

  • Replication for backup purpose

    We are using Berkeley db on server in a highly concurrent environment. The no of requests per second is huge and we can't afford to lose any committed data. We have 1 master where all updates will go and that's the box that is serving all the UI requests. We are planning to replicate the bdb to other backend nodes for readonly purposes. We don't want any node other then the UI node to become a master(basically no elections otherwise the writes would become complex). The question I have to all Berkeley db gurus is that we are planning to have a standby replica of Berkeley db just for backup purpose, currently we checkpoint every 2 minutes and rsync to another machine but even loss of data for 2 minutes is catastrophic. We had to do a catastrophic recovery once already and lost some data.
    Is this replicating to a standby node for recovery purpose a reasonable strategy that people are already using or am I missing something completely and there is a better way to do this?

    Hello,
    This question generates more questions than anything else. What version of BDB are
    you talking about?
    Your question also implies, since you talk about
    checkpoint and losing 2 minutes of data, that you are running with DB_TXN_NOSYNC. Running
    with DB_TXN_NOSYNC opens up durability vulnerabilities that must be weighed against
    performance. Using replication is possibly a reasonable approach, depending on details
    of course. You likely want to investigate memp_trickle. However, only you can weigh
    the durability versus performance tradeoff of NOSYNC, but realize it is a tradeoff.
    Or you can consider running with NOSYNC on the backup, but not on the master.
    That has different performance tradeoffs depending on acknowledgements (see below).
    First, if we assume you are using the Replication Manager API, then you need to properly
    set your priorities on the backup(s) to 0 (you talk about multiple nodes sometimes and
    a single backup at other times so I'm not sure how many replicas you intend). However,
    by setting priorites to 0, then you need to understand what your strategy is for failure.
    What steps do you intend to take when a failure occurs (or when the former master is
    restarted)? What about failure of the backup?
    For instance, if your replicas run with DB_TXN_NOSYNC, then the only way to
    synchronize a rebooted former master that lost its data is to have it join the
    group, and some other site become master so that it can synchronize with that
    master. On the other hand, if you manually copy the databases and logs from
    the replica to the failed master site, then you lose any txns, pages, and log records that
    are still in the memory buffers. So there are several combinations of tradeoffs
    there. If you run synchronous txns on the backup, manually copying the
    databases and logs should work, but you have the performance impact that
    will delay the replica's acknowledgement.
    If you're using Replication Manager, what ack policy are you using? The answer depends
    on how many sites you have and how you configure them. If you have more than 2
    sites total then you likely want QUORUM policy. If you only have 2 sites, then you may
    want the ONE policy or something else depending on whether you have set the
    2SITE_STRICT flag or not.
    As I said, there are a lot of tradeoffs in many places and BDB gives you the
    flexibility to make those tradeoffs. But you need to understand your requirements
    and understand what tradeoffs you make and their impact.
    Sue LoVerso
    Oracle

Maybe you are looking for

  • IMessages not showing on iPad

    I'm trying to understand why iMessages from my iPhone 4 won't show up on my iPad 2. I have Receive At set to the same E-Mail address. The only thing I can think of is that the person sending it to me is using my phone number as the recipient and not

  • Error in Creating Project in ESS.

    Hi All, I have a requirement where in , there is change in the personal data screen of ESS package. We configured the NWDI server. But when i am trying to extract the ess/in/pdata dc from the ESS SC. It is failing during the project Build and throwin

  • How to fix the length of the output file

    hi, I have 1000 records in my source table which need to be inserted into file. But the length of the rows in my file should be only 500 i.e there should be two files getting created for this 1000 records i.e my each output file should contain only 5

  • Oracle direct path read IO size

    Hello! I am confused a little with IO size. I am running 11.2.0.3 on Oracle Linux x64 6.2. Block_size=8K, MULTIBOCK_READ_COUNT=128 Database is Single Instance and is using ASM grid. ASM AU =1M As a test I am running a simple query against large table

  • Value based auditing

    Hi, what is the best way to implement value based auditing? Usually you have to use triggers, but can I also do it with data changed captue? Peter