Jump to content

Adyn

Member
  • Posts

    29
  • Joined

  • Last visited

Awards

This user doesn't have any awards

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Never doubted that you would not have to be honest as that number of hosts does suggest you would have planned for power, however out of interest when you have grid failure is your contingency to run forever off diesel until the power is resumed. Its just when we modelled it we looked at a significant number of traffic accidents in the region which meant getting a refiling tanker to the locations could not be guaranteed so we would look to move sites after a couple of hours?
  2. From a Honeywell tech document for one of their windows 2000 based products in this case for CCTV. "The System shall only require a single security key dongle to be present on the database server for the System to operate. Security keys shall not be required at the client workstations. The System shall allow a user to read the information that is programmed on the server security key dongle. The System shall support export of the information using the ‘Export Dongle information’ button, which shall allow the user to forward to the integrator when upgrading new dongle features." So in this case (and it is a random example) you need a security key in order for it to work and you have to export this information to another key which would go back to the supplier if there was a change to the system. So in this case and there is no real way of knowing if the security key could detect a change to the underlying hardware (which would happen both if you re-platformer or virtualised) you may then have to get an Export dongle get it updated (however that happened) and reattach it to the whatever you have moved the machines onto. Then this becomes from a VMware perspective doing either serial or parallel port pass-through and that is the part that is not always straight forward as who knows how they decided the system would talk to the security key they may have created their own method (as I have seen in the past) that just does not play well with others.
  3. True however Do so Yes there is no issue with the OS and yes you can virtualise any flavour of OS without a problem, but the question is not about the OS its the application that runs off it hence the reason to speak to a To ask the question. as your would not work if the power station they are powered from stops because your. needs a hardware dongle.
  4. I would tread very lightly with this for several reasons. Firstly, and easiest to cover is the licensing if it is the original supplied hardware it will most probably have OEM versions of the OS you can get around this if you are licensed for a current version of windows you can legally run older versions. The application side of licensing is where things get tricky many older control systems software were licensed by a physical component either permanently attached (serial or parallel connection) or required for installation in which case it is either going to be impossible or seriously challenging to get this to run virtually or even to re-platform on new hardware. You can check any documentation to tell you this. Saying that you do have a system that is out of support from an OS perspective so the ideal place to keep legacy systems (if you can P2V them) is in a virtual environment. In this case I would go for VMware ESXi as oddly enough windows 2000 is still a supported guest from a hypervisor perspective. Yes you may get resistance from the manufacturer / supplier which is why what I must strongly suggest is you reach out to them to ask what they would advise you to do in this situation.
  5. In you price point I would look at HP ProLiant Micro servers you may have to do a bit of hunting to find one at your price but found a couple (http://www.ebay.co.uk/itm/HP-Microserver-G7-N40L/122216989538?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D38661%26meid%3D8a769f4f63cd43f785e349bbd2ad4f66%26pid%3D100005%26rk%3D1%26rkt%3D6%26sd%3D172406554249) You may not get one with the total amount of storage that you are after but it does give you 4 drive bays (more if you don't mind using the CD bay). It should be noted though that at this price point everything is going to be older kit but at least with a microserver your power bill will not be through the roof.
  6. Ok I just read what you are wanting do use for a client and it makes more sense that you are wanting to use what I would call proper thin client. I would ask however as Citrix sits on top of Microsoft remote desktop for ten users I would not see that much of a benifit adding Citrix into the mix unless you require a specific feature. With regards to licensing with just Microsoft Remote Desktop Services (RDS) you need a license server (which could sit on a server tasked for other things) but when you add Citrix on top that requires its own license server as well. Also reading that you were wanting to use Xenserver as the hypervisor you are in effect talking about putting a virtualisation product on top of a virtualised infrastructure and while with ten users you are unlikely to have issues when you start adding more you would have to be careful with your storage IO as I have seen more than a few of these go horribly wrong if the correct workload planning is not done.
  7. What type of Citrix are we talking about is it the original thin client variety or the VDI version as the requirements for both are completely different. If you are thinking about using it then it would help to understand this.
  8. I would agree with this to be honest as this is about as cheap as you can get although you may save more with a static shelf. But if it is more down to cost you could argue that you already have the rack the cheapest option is to place whatever it is on the device below just playing devils advocate.
  9. There are two types of server clusters distributed and attached. Distributed clusters take a workload activity (task) and spread it between multiple hosts they generally have control node(s) that allocate the activity and collect and serve up the results. These are typically compute intensive and spread the workload in order to achieve their results. If an individual node does not respond with the work package, then this is served to another node in order to carry out that task. Attached clusters share storage between the nodes (and when I say attached it does not necessarily need to be attached to the devices it can be over a network but then storage area networks (SAN) are networked so the analogy is just for understanding). Attached clusters share access to a common storage medium so guest machines running on the cluster (or guest processes depending on the type) have their activity visible to multiple hosts. As Andster29 pointed out this gives a measure of redundancy as if a physical host fails the other hosts can pick up the workload with only the information running in the physical memory and CPU cache being lost. As for how are they made it really depends on what you are wanting to make to either distribute workload or improve availability in order to answer that and that would be an entirely different question.
  10. The first rule of server club is there is no such thing as servers. Honestly forget about servers there is no difference between a “PC” and a server but for some reason they have this almost mystic quality. Yes, the pictures look cool but if you boil it down there is very little difference. If I am being honest I way more prefer building custom PC’s than servers, they are way more user friendly and you can get almost every possible configuration into a desktop form factor. I have servers at my home (and more than the other half knows about) in my basement and I never turn them on because they are so loud and use so much power. Yes, you can get cheap ones on eBay etc. but odds on they are at least 3 if not 5 years old (because they are classed as assets and from a financial perspective are kept by organisations for that long due to depreciation). If you want to be a hardware engineer, then it is kind of pointless learning old hardware but if you want to learn about the server software side of things you can build a home lab that would be way easier I made a post about such things. But in general a server is just a PC without anyone sat at the screen (if it has a screen).
  11. Have a look at http://www.expta.com/2016/05/expta-gen7-home-lab-server-builds-and.html it looks exactly what you are after. I actually built one of the older ones and it would happily run six virtual machines. The only thing I would suggest is that you use a bigger case as it makes it much easier. But have a read through the link as its written by a Microsoft MVP and it explains the rational about the components but the most important factor is the motherboard to ensure you can fit enough memory. Then look at the workload you are wanting to deploy and let that dictate your CPU choice. If you then put a decent graphics card in it would happily play games. One final point to ponder you are saying that you would be using Microsoft server products unless you can afford to buy them I would run a Windows 10, 8 with hyper v (or another virtualisation desktop product) and install your server virtual machines from Microsoft 180 day evaluation versions, anything you want to use as a permanent setup use an open source OS. Its just if you setup a cool home development environment and want to use it to demonstrate self development for an interview one of the trick questions will be how is it licensed ;). Oh and you can run a hyper visor (Xen, ESX etc) on a hyper visor think virtualising a virtualisation product and then you will find your home lab can do some pretty cool things.
  12. Filled it out however you may struggle to find people who have suffered electrocution and able to fill out forms as electrocution is death caused by electric shock, electric current passing through the body. Also being in the UK our plugs already have the anti electric shock features built in so not sure how applicable this is but hope it helps.
  13. If you are looking for something specifically to look good on a resume I would look at machine learning and “big data”. More specifically looking into neural networks. While it does on the surface look a daunting subject there is plenty of information around if you tied that with as has been previously mentioned suggestion with raspberry pi and Arduino you could have a project that covers off real applications. That would give you for a resume perspective a project that covered a significant number of different but in demand areas which would look very good on your resume and show that you were willing to educate yourself with in demand areas that have a significant skills shortage.
  14. Think waves on a beach while you have peaks and troughs they are different heights and depths, that’s analogue. Anything with waves is analogue so think sound, electricity etc. The problem with waves is they are not static so if you were using electricity for a control and you wanted to sample for a certain voltage to use for controls you would get multiple readings at random when you wanted a single value. Hence the requirement for digital so you can have control this is achieved by using a transistor to measure a voltage over a period (think as small as you can) to determine if that signal is a 1 or 0. So now that you have a defined digital signal you can use that to carry out calculations store data whatever you can think of. Now this is the bit audiophiles do not like you can reverse this process to get an analogue output but it is always a representation of the original source. Now you can do things to decrease the sampling time to get a more authentic signal but if a digital process is used anywhere then that becomes a representation of the orginal.
  15. Yeah not sure how active directory is working without DNS
×