|
What's the best solution for fast hard disk access in a home PC? From what I've read it looks like multiple drives in RAID 0 configuration would be best.
|
|
|
|
|
Personally I'm not sure there's much benefit for RAID 0 in a home PC. You only get any benefit from striping if you have multiple concurrent disk accesses which target different disks, which tends to be the case if you have random access I/Os. The vast majority of I/Os in most systems are sequential - we've all written the code, open a file, read it to the end into memory data structures, close the file. When finished editing, open the file for writing, spew out the data structures again, close the file. All sequential.
With sequential I/Os, each drive in a striped array (RAID 0 or 5) effectively becomes active in sequence as the file pointer moves across the disks. Only if you have a lot of concurrent I/Os to different files might you get a benefit, but even then you don't get a lot of advantage.
RAID 0 has the obvious greater risk that if you lose one of the disks, you lose the volume. The probabilities are greater that you'll lose at least one of the disks than if you were to lose a single standalone disk. Compensating for this by using RAID 5 has a large disadvantage in write speeds.
The best place to start is to get disks with high spindle speeds. The higher spindle speed increases maximum bandwidth from the platter, and slightly decreases access times as the desired block will appear under the head more quickly, on average. At the very least get 7200rpm drives, 10,000rpm SATA drives are also available for a premium and generally have lower capacity than 7200rpm drives. (750GB 7200rpm drives are available for about the same price as 150GB 10k rpm drives).
Because of the varied access patterns for different applications and the different loads to which systems are put, it's really difficult to tune the disk subsystem to any degree for a home PC. You might get most benefit from simply throwing lots of memory at the system, so that more data is cached.
DoEvents: Generating unexpected recursion since 1991
|
|
|
|
|
Recently, I've found myself running out of contiguous storage space. (I've got a lot of small drives installed over many computers. Some of them are even lying in my cabinets.) I decided it would be best to have a centralized (or at least portable) storage medium to stuff my files in. I've been hunting around the internet scrutinizing many products being advertised.
My requirements for the products were:
- Does not require installation of packaged software
- Can be accessed using one, any combination, or all of these: Firewire, Ethernet, or eSATA
- RAID-5 support (and by extension, it holds at least 3 hard drives)
- Supports NTFS
- Internal use of SATA preferred over IDE
- Massive amounts of storage (3 to 4 TB preferred)
- Something that is asthetically pleasent (e.g.: no giantic logos or company name carved in cursive…)
iOmega's 2TB Power Pro[^] seem to have hit most of the points. I'm considering getting one.
Does anyone have something that they think is better?
Thanks in advance…
ROFLOLMFAO
|
|
|
|
|
Well, I use a Seagate FreeAgent Desktop, but maybe you would like to look at the pro version of FreeAgent[^] They come in max size 750GB, but the hard disk and controllers are really good. (atleast in my lesser-pro version  )
//Johannes
|
|
|
|
|
Thanks. I'll look into it.
ROFLOLMFAO
|
|
|
|
|
Ri Qen-Sin wrote: iOmega's 2TB Power Pro[^]
RAID 5+hot spare Isn't this just RAID6?
--
Join the Campaign to Help Stamp Out and Abolish Redundancy
The preceding is courtesy of the Bureau of Unnecessarily Redundant Repetition Department.
|
|
|
|
|
Nope. RAID 6 protects against double disk failures. If a RAID 5 + spare array, however, would fail if it were to have a double drive failure (failure within the time the first drive fails and the rebuilding process).
Standard RAID Levels on Wikipedia[^]
ROFLOLMFAO
|
|
|
|
|
Hi guys, I just need some general info, and I'm hoping someone can point me in the right direction.
I have an old (like 6 years old) ink-jet printer, and I want to be able to control the page feed so it can print in both directions, instead of always feeding the pages in the same direction.
Is this even possible? If so, how? I'm not looking for a detailed solution, just general information on how I should go about doing this, i.e. do I need to make my own print driver or hack the the hardware or what?
|
|
|
|
|
if what you need is the capability of both portrait and landscape printing, that is
provided by Windows itself. You can control it through the Page Setup or Printer Setup
dialog available in most apps.
Example: MS Word, Page Setup, Paper Size, Orientation
if you need to add landscape printing to your own app, it basically means tell the print
dialog that's what you want, Windows will do the rotation for you, the only thing you must
take care of is applying the page bounds that you get.
Luc Pattyn [Forum Guidelines] [My Articles]
this months tips:
- before you ask a question here, search CodeProject, then Google
- the quality and detail of your question reflects on the effectiveness of the help you are likely to get
- use PRE tags to preserve formatting when showing multi-line code snippets
|
|
|
|
|
Thanks for the reply, but that's not what I meant
|
|
|
|
|
Some old dot-matrix printers with tractor feed could move the paper in both directions. I don't believe that you can do the same with a sheet-feed printer, the paper would probably jam and ink would smear even if you could order the printer to reverse the paper direction.
|
|
|
|
|
Thanks for the reply
Paperjams and smearing won't be an issue because I'm completely rebuilding the housing structure. I just need to know what approach I should take for reversing the paper direction, if it is possible.
Is it the print driver, commands sent to the printer, or do I have to hack the hardware itself?
|
|
|
|
|
You will have to hack the hardware, and since its probably a simple DC motor that pulls the paper through (atleast on the old ones ive taken apart) then you should be able to reverse the polarity. If it has a stepper motor then im not sure if you can move the wires around to reverse it or not.
You will have to push the paper in backwards though...  and it will probably try to push out any paper lying in the paper tray too 
What are you going to do this for, it seems funny, a practical joke??
//Johannes
|
|
|
|
|
I'm making a 3D printer, like a Zcorp printer. So, the first step is to turn my old printer into a flatbed printer, and ideally, I would like to be able to print in both feed directions, rather than zipping back to the start position each time, if you know what I mean.
|
|
|
|
|
mike7373 wrote: Is it the print driver, commands sent to the printer, or do I have to hack the hardware itself?
All of above, and the firmware in the printer's embedded controller. Essentially you would be building a new printer that uses some of the components of the old one.
|
|
|
|
|
I just read the whole thread
Why don't you plant the print-head on a flatbed-scanner chassis? Or a plotter?
If you want to print in 3D, you'll probably want to go with lowering the target instead of raising the "printer assembly", and a flatbed scanner usually does not have PCBs at the bottom (unlike an inkjet).
If you need to go with an inkjet printer alone, find one which also can print CDs and takes them from the front side. That way, you might get a printer which can a) reverse and b) might not jam when feeding backwards.
Cheers,
Sebastian
--
"If it was two men, the non-driver would have challenged the driver to simply crash through the gates. The macho image thing, you know." - Marc Clifton
|
|
|
|
|
Addendum:
You'll really want to go with the scanner, though. You just dont want the target to move, trust me. 
Cheers,
Sebastian
--
"If it was two men, the non-driver would have challenged the driver to simply crash through the gates. The macho image thing, you know." - Marc Clifton
|
|
|
|
|
Thanks, that's a great idea.
I had planned on lowering the target, as you suggested, but I never thought of using a scanner, and I just happen to have an old scanner lying around too 
|
|
|
|
|
I am working on an application service that will be running on a PC with no monitor attached to it. Is there a way to force the hard drive activity light to blink (maybe like 1 second on, then 1 second off) so that we can tell whether or not the service is running without attaching a monitor? The normal disk activity that occurs while the service is running happens so fast that the light just flickers so dimly that it is hard to see if it is even on.
Thanks for any help, or pointers to resources that may be useful. The only things I could find from searching are people asking why their activity light constantly blinks.
--
Marcus Kwok
|
|
|
|
|
I would not think that there is any way to directly control weather the HDD LED is on or off, but simply replacing a file with a large amount of data at certain intervals would work. Or if you are lucky your computers still have their PC speakers installed, meaning you could use Beep(int freq, int duration); (duration in ms)
Or maybe you could use inpout32(); to control the binary outputs on a LPT/printer port, and then connect a LED and resistor to make it blink... [^]
//Johannes
|
|
|
|
|
Johannes,
Thanks for your reply.
Johpoke wrote: simply replacing a file with a large amount of data at certain intervals would work.
This probably would work, but my concern with this method is that it might contribute to the hard drive failing sooner than otherwise. In this application, the device is intended to be in service for many years with minimal user intervention. For the units that we are replacing with this new unit, the hard drives were the most common point of failure.
Johpoke wrote: you could use Beep(int freq, int duration);
This method might be a little too intrusive/annoying for the people working in the room.
Johpoke wrote: Or maybe you could use inpout32(); to control the binary outputs on a LPT/printer port, and then connect a LED and resistor to make it blink...
However, this idea is very interesting. Thanks for the pointer!
--
Marcus Kwok
|
|
|
|
|
ricecake wrote: This method might be a little too intrusive/annoying for the people working in the room.
I was thinking the room might be like the server rooms ive been in. If that where the case you would have to put you ear to the case to hear the beeping. (or *overclock* the speaker )
Glad I could help 
//Johannes
|
|
|
|
|
Well, it's kind of like a server room in that there will be a large amount of equipment in there, but I believe people will be working in there for a significant part of the time as well.
--
Marcus Kwok
|
|
|
|
|
Well, the keyboard LEDs would work. You can control those with an interrupt, IIRC. Did that in TurboPascal once.
As for the HD-LED: difficult.
Cheers,
Sebastian
--
"If it was two men, the non-driver would have challenged the driver to simply crash through the gates. The macho image thing, you know." - Marc Clifton
|
|
|
|
|
I don't think you can mess with them that way in a modern OS. My turboPascal keyboard light toggler crashed under NT4(3.5????). There's got to be an API way to do it though.
--
Join the Campaign to Help Stamp Out and Abolish Redundancy
The preceding is courtesy of the Bureau of Unnecessarily Redundant Repetition Department.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
