I ran into an issue at 62,000 words on the book manuscript, being that MS Word struggles to render any pages that have embedded images, even when I’ve been careful to render the image at the lowest acceptable quality of image resolution – the problem was so bad, that the entire computer would seem to have crashed for a few minutes, and which is not acceptable given this is a reasonably powerful machine that should have no such issues with a simple display of text, formatting, layout, and the occasional image.
NOTE: The terms used below are explained further down ( after the review ), if you’re interested in a quick lesson about system configuration for your own home/office computers.
The other issue I experienced was that being a newer CPU ( Central Processor Unit ), Microsoft and Intel have unnecessarily colluded to prevent the bare-metal installation of Windows 7, in order to force people into the Windows 10 upgrade – but my impression of Win10 so far is that it is an absolutely unstable, unreliable, and poorly designed piece of utter garbage, which breaks many principles of best design practice, and has demonstrated several serious and troubling stability issues.
So being frustrated with the seemingly never ending task of reconfiguring computers, I came up with a solution after a little research, but it requires some extra pieces of hardware, and I’ve never done anything quite like it before.
system requirements:
What I’m looking at doing is as follows:
- Install a base Linux OS ( Operating System ) on a RAID5 ( Redundant Array of Independent Disks ) array;
- The RAID will consist of 3x 512GB SSD hard drives;
- This requires one additional SSD ( Solid State Drive ) to make the set;
- This will provide both redundancy and speed;
- Create a shell script that will shut down the Linux desktop and other non-essential functions;
- This allows me to use the desktop if I want, but without having to boot into a different install;
- Create a Win7 VM ( Virtual Machine ) with GPU ( Graphical Processor Unit ) passthrough;
- This allows the VM to access the graphics card directly;
- This also allows near bare-metal install performance of the VM;
- Thus I can bypass the way they’ve locked out Win7 installation;
- Other VMs will be used for development and testing;
- Cloning and snapshots of VM states provides additional backup/redundancy;
- At a later date, a 2nd 2TB hybrid drive will form a RAID1 ( mirrored redundancy ) data store;
- I will get the base Linux OS to offer RAID1 partitions as a NAS ( Network Attached Storage ) service;
- I will switch from MS Word to trying other software like CeltX and Scrivnr ( for writing );
- I will install a GitLab instance in the base Linux OS for collaboration and development;
For those of you who aren’t techies, think of it like this:
When you have a complex environment – but minimal resources – you’re forced to try to fit a bunch of potentially conflicting software installations on the same machine ( physical hardware ), but this tends to lead to more technical issues, which then requiring solving. By creating a machine that uses VMs, you can tailor each VM to a particular type of work, and thus create a demarcation of functionality, which can prevent some of these conflicts occurring, and thus save you a lot of time, frustration, and risk of data loss.
end goal:
What I’m trying to create is more than just a high end system with built in redundancy, and more than just a workstation replacement laptop, but a single portable work machine capable of acting as a portable home-office network of machines.
My old laptop ( after a few repairs and replacement parts – the hard drives were both dying, and the previous GPU died and required replacement ), is eventually going to become a purely entertainment machine ( running a TV and sound system ), and also a few other basic tasks, such as being a printer server for the home network ( the old printer I have does not offer a network connection of its own ) … if possible also a scanner server for the LAN ( Local Area Network ), and an internal web server for some LAN services and development / testing functions ( just to take a little load off the new machine ).
metabox review:
In this particular case I found an Australian based retailer Metabox, whom provide a workstation replacement laptop that has a total of 3x 2.5″ HDD ( Hard Disk Drive ) bays, and 2x 3.5″ HDD bays. I could not afford to get everything I wanted up front even after 2 years of saving, plus a large donation and compensation for lost property in the floods following cyclone Debbie, but I was able to get enough of it, so that with a little more saving I can purchase 2 more hard drives ( 1x SSD and 1x hybrid ), then at a later stage I can get a bit more RAM ( Random Access Memory ), and thus create a single physical machine with multiple VMs available for specialised usage requirements – even running multiple VMs simultaneously.
I would like to say of this company Metabox, that in general they’ve been very helpful and patient with my support requests. They only offer support for Windows 10 on these machines, but having explained my needs and intentions to them, they’ve been quite willing to help within reason, so that I can get the environment I desire, and which suits my purposes.
The manuals that come with the laptop are ( I think ) a bit rudimentary and generic, so really needing a lot of rewriting ( or an additional manual written for technical users ) – but for a non-technical user I guess they’re fine. For the moment – if you’re a developer or techie – you have to contact them to get the additional information you require, in order to do really basic upgrade and maintenance tasks, such as replacing a hard drive or adding RAM.
The manuals are generic – meaning they’re written by someone else ( no Metabox branding ) – and my understanding from another review is this is a consequence of Metabox being simply a branding exercise in many regards ( an overseas manufacturer makes the housing, but Metabox Australia then configured and supports systems using that housing ), with their branding on the surface.
The manual itself and warranty conditions currently imply ( state ), that your warranty is void if you open the housing yourself, but Metabox contradicted this in correspondence with me, where they stated that for some simple tasks, they allow customers to do their own upgrades – which is likely a consequence of the fact their manual and documentation is written by that generic manufacturer overseas, whom probably supply to multiple rebranders/resellers ( like Metabox ), and then it’s up to the reseller to override ( or not ) those generic terms and conditions. On which topic my feedback to Metabox has been, they should rewrite their materials, since this may be good for non-technical users, but if they’re marketing themselves to gamers, they need to allow for the fact that many gamers are techies, just as many techies are gamers, and the hardware specs they sell are great value for money for both markets, so they should make an effort to present better to the techie market, and remove any doubt or ambiguity as to the terms and conditions of the warranty – as many users like myself, don’t want to call tech support unless it’s an absolute emergency ( which an upgrade shouldn’t be ).
In all cases I would advise you to ask them first to be sure of what you can and can’t do with the warranty conditions, as this appears to be a slightly informal situation at present, and what they said to me in correspondence may not apply to all users, especially if they’re not confident of your technical ability ( which is fair enough if they’re going to pay the cost of repairs on warranty ).
In much the same way, they don’t offer much in terms of customised software, so it’s a generic housing, with a generic manual, and a couple of bits of Win10 software ( CPU/GPU/fan control etc. – for things like CPU/GPU frequency boosting ).
The big advantages are the fact they are far better value for money than someone like Boxx, and with local Australian service/support. By comparison, Boxx manufacture all their own housings, they custom design cooling systems, write their own documentation, and they specialise in extremely high end machines for 3D CAD/CAM/CAE designers and engineers across the entertainment and manufacturing sectors – including specialised render farm machines … HOWEVER: while their designs, engineering, and support may be better, they’re in a different time zone ( though they do have local support partners ), but more importantly – especially given the currency difference – the same core hardware will cost you up to several thousand more from US-based Boxx vs. Australian based Metabox.
Removing the casing is not made to be an intuitive task – after removing several screws, I was still unable to remove the casing, having tried lifting, sliding, and various other actions, but not wanting to risk damage by forcing it. By comparison, the Alienware laptops are quick to access internal components with only 2 screws to remove the casing, and come with a technical manual with detailed diagrams for removing and replacing every core component … which makes sense, as the last thing you want to do is ship your computer off somewhere for something as quick as this, wait for them to do it, then ship it back – nor do you want to have the cost and inconvenience of in home-office servicing by one of their technicians, if you could just pop the lid and get it all done any time of day or night, in just a few minutes, at no extra cost, and get back on with the job of whatever you were doing.
If you’ve got the new nVidea GTX-1080 GPU ( Metabox offers 2x 1080 in SLI ), the fans are a bit noisy, but you do get used to it.
recommendation summary:
For a non-technical user, all the above is probably wonderful, especially if you pay for the on-site service, but if you don’t buy computers to pay others to work on them ( like myself ), it’s a pain in the arse, but at least a pain they’re willing to negotiate around. They have been quite helpful ( if sometimes a little slow in response ) in providing the assistance required to go beyond the aforementioned limitations.
Artists and designers who use software like Adobe, Corel, Maya, Revit, SolidWorks – or any other design / engineering software – would find themselves well serviced in value for money by Metabox.
CAVEAT: having sold SolidWorks myself, I would recommend a larger organisation should only compare Metabox vs. bigger companies Dell, HP, or Boxx, by first discussing with Metabox your precise service requirements. The rationale being that when you’re paying a service an support contract for many workers – possibly in multiple offices and time zones, and where critical contact-sensitive project time-lines are involved, with direct impact on your revenue and business relationships – you’re better off negotiating a custom support contract, with delivery guarantees, even if that means you’re paying more and the company in question actually has to hire and train new staff just to deliver it. In fact I would argue you should do this anyway, even with a big company like Dell, because IT management often sees support contracts as pure gravy, and while they’re perfectly capable of delivering better quality support, they often don’t.
terminology and further explanations:
cpu:
Intel and Microsoft have colluded on the latest generation of CPUs, not to offer a driver for Win7, and thus force the upgrade to Win10. I did however find a community open source project mentioned in a forum, to provide a CPU/chipset driver and thus allow for it to be installed, so if that’s true, then it confirms that it was possible and should have been offered by Intel and Microsoft themselves ( its still x86 architecture, so there’s no reason for it not to be done ) – which maybe they do for very big paying corporate clients (?), but even if so, there’s nothing otherwise available from them … but the instructions on how to use this driver was ( I thought, like so many other technical articles ) poorly written, and seemed quite a convoluted process; not that I’m expecting it to be easy, but it should be better described and detailed.
The newest of the new generation of CPUs now have more than 4 CPU cores, which means as the price comes down, and as these desktop CPUs become more readily available in workstation replacement laptops, you’ll be able to do what I’m doing with a host OS serving up tailored VMs, such as to have multiple simultaneous customised client VMs, each more powerful than a present day desktop of the old 4 core CPUs – an entire VM based development LAN on a single physical device.
linux:
Linux is an open source Operating System ( OS ) originally written by Linus Torvalds, based on Unix, and which is well known for its efficiency, security, and stability, particularly in the server environment – even Microsoft runs Linux severs for many things, despite producing their own Windows Server OS. Linux has a number of options for creating and running VMs, it can be installed with or without a GUI ( Graphic User Interface ), and thus can be entirely operated by the command line.
ssd:
An SSD is a hard drive that functions by RAM ( Random Access Memory ) rather than ROM ( Read Only Memory ). In a RAM drive you can instantly and directly access any memory address, limited only by memory ( access ) and BUS ( communications ) frequency; whereas in ROM you have to read through the drive sequentially to get to the start point of the data you’re actually after. The difference between RAM chips ( operational memory ), versus a RAM drive, is that the latter is persistent data, whereas the operational memory is gone when you switch the computer off.
hybrid:
A hybrid drive combines the old disc spindle with a small SSD, thus it is cheaper than an SSD, can offer more memory at a lower price, but is faster than a purely spindle based ( ROM ) HDD.
vm:
With virtual machines, a base OS is installed bare-metal on the physical hardware, and acts as a host to potentially many VMs, usually by offering up a virtual interface to the physical hardware. So the parent ( host ) OS deals with the hardware more directly, but creates a virtual version of each device, then the guest/client VM’s OS – which might be a completely different OS to the host OS – speaks to the physical hardware only via those virtual devices, so that the host can schedule its own hardware requests alongside any client ( VM ) hardware requests, without conflict.
A VM can be cloned ( a full backup of the VM in a particular state ), and can store “snapshots” of its various states, such that you can revert to a previous stable state if something goes wrong, and transfer a cloned VM from one set of physical hardware to another.
So if you’ve ever been in the situation where you’ve had to format your hard drive, reload the operating system, and reload all your applications and data, a VM can prevent you ever needing to do this again – and a VM can be conveniently copied ( cloned ) as a backup.
raid1:
RAID1 is a mirrored array, requiring 2 drives, so every bit of data is written to both drives ( taking more resources and thus slowing down write-speed slightly ), but can also therefore be subsequently read from both drives simultaneously ( faster ) … and more importantly, if one drive dies, you haven’t lost any data, and you can keep working until the drive is replaced ( at which point the RAID hardware or software controller, will rebuild the RAID for you ).If you only have 2 hard drives and no other backup, a RAID1 is a good option, on which you’ll install your OS, applications, and store all your data – but from your user perspective, the two physical hard drives will appear as 1 single logical drive.
raid5:
A RAID5 array requires a minimum of 3 drives, and what the RAID controller will do is write a variable data-bit pattern of { data1, data2, parity } – ie: the “parity bit” will change position for each sequence of 3, as to which HDD it is written on – and the value of that parity bit is determined by algorithm such that, in the case of hard drive failure, the RAID controller can figure out from the other 2 bits, what the value of the 3rd bit in the set should be. The consequence of this, is that a RAID5 array can withstand the death of 1 Drive in the array, but continue functioning ( albeit at a slower speed, with resources being consumed to calculate the value of the missing drive, until it is replaced ).
The value of a RAID5 is thus both speed and data backup/redundancy, as it can read and write to all drives simultaneously, but losing only 1 drive of storage capacity for the parity bits, and losing only a small overhead for calculation and read/write scheduling of those parity bits, but still faster than a RAID1.
A RAID6 by comparison requires a minimum of 4 drives, in order to write the set { data1, data2, parity1, parity2 }, which can then withstand the simultaneous death of 2 drives without ceasing function.
Hence this is the great advantage of Metabox offering a total of 5 bays in the model I purchased, giving me the capacity for a RAID5 operating system partition, and a RAID1 data store partition – where of course I can therefore swap in/out VMs between the RAID5 ( operation ), and RAID1 ( backup ).
NOTE FOR BUYERS: the average graphic artists and desktop publishers might struggle to understand all this, but the point being – without purchasing an additional backup machine ( such as a NAS ) – you can have both speed and backup on a single mobile device, which is a little heavy, but it’s like being able to take your entire office with you. If you can afford to do so, and you don’t want to carry that weight for short trips, then you could get a slimline laptop for day outings ( eg: seeing a client ), which contains just a bare minimum install of work related software, therefore also doubles as a machine to handle largely entertainment / recreational software ( like iTunes ), which can be a good idea, since I’ve often found software like Apple iTunes is poorly written for PC architecture and the Windows OS, and some of the bugs in it have been around for a decade with Apple doing little or nothing to solve it – so I’m inclined to keep such software off a work machine, and which I’ll be able to solve myself by putting all that on a separate VM, or on my old laptop.
gpu:
Many cheaper computers have only simple graphics processing on a chipset built in to the motherboard, with support from the CPU, so if you’re doing 3D design work, they can really struggle – whereas a proper and more powerful dedicated GPU device offers a huge amount of resources and grunt for graphics tasks, freeing up your CPU cores or the actual work you’re trying to do.
passthrough:
GPU passthrough is where a host OS offers direct access to the GPU for client VMs, which ( apparently – I haven’t done it yet ) can equate to 95%+ performance of VMs versus a bare-metal installation of whatever the client OS happens to be.
nas:
Network Attached Storage is kind of like a dedicated OS that deals primarily with allowing computers on the LAN to access a storage partition, which means it can also apply access rules, and it can also offer related services, such as FTP ( File Transfer Protocol ) access, or a local web server. Companies like QNAP and others offer dedicated NAS devices, and which usually run from a RAID5 or RAID6 array, so you can buy them any size you want.
NOTE FOR VIDEO EDITORS: if you don’t have a NAS, you should.
gitlab:
GitLab is an extension of Git ( the extremely efficient and decentralise, open source file version and revision control repository system, developed also by Linus Torvalds – see Linux ). In simple terms the difference being that Git is a purely command line utility, whereas GitLab adds a web server and related services plus new functionality to manage teams, projects, and various other things.
Thus GitLab – the open source community edition – or something like it, will be at least the inspirational basis of the core components of my proposed Project Collaboration Development & Resource Allocation Framework of the Open Empire vision.