[Home]TechSupport

ec2-34-205-246-61.compute-1.amazonaws.com | ToothyWiki | RecentChanges | Login | Webcomic

A hub for techie questions that don't already have a specific section, and links to specific techie sections.




Start the services console: %SystemRoot%\system32\services.msc /s


Not really tech support I guess but I can't think of anywhere else to put this:

Does anyone know if someone fairly inexperienced in wiring (i.e. me) could rip apart a USB cable and wire it up such that a bunch of LEDs flashed when 'data' passed along the cable?  The main requirements being that the USB cable would still work and not damage the things on either end of it.  Failing that, would it be possible to just rig a couple of LEDs to light up when power was going through the cable? --K

It's certainly possible.  However you've stated two seperate aims here.  Do you want an LED when data is being transferred, or when the power is on?  It is pretty much trivial to stick an LED on the power lines - but the data lines may be less happy about it.  [Diagram] - might be helpful for you to identify which the power line is.  --Vitenka

Sorry, thought it was clear...  I'd like to have the LEDs reflect the data activity but if that would be too complicated I'd make do with the LEDs reflecting the power status.  The diagram mentioned is for the wiring between the motherboard and the USB port.  Is there a trivial mapping between this and the USB cable itself? --K
http://www.connectworld.net/usb.html gives cable colours and pinouts for the connectors. I've never taken a cable apart, but I suspect the cable colours might not be strictly adhered to, so it might be worth it pulling apart a connector, working out which wire attaches to which pin, looking the pins up on the diagram and going from there. What is it you're gonna use it for? If you want to work out whether a usb socket is carrying power, it might be slightly less painful to pick up an unpowered USB hub for a couple of dollars from a wal-mart and see if its LED lights up when you plug it in. Also check out [this] - clearly it's possible, so there's probably circuit diagrams on an overclocking site somewhere or other.. - MoonShadow
Roughly speaking, inserting [one of these] into [one of these] and, if the truly desturbing is required, making the eyes flash on data access...  Probably won't happen, not least because of the price of the first, but it would be amusing if it did... --K
If you're spending that much anyway, you could save yourself a load of bother, get a USB extension cable and one of the flashing cables I linked earlier, hack both apart - the flashing cable next to the flashing plug, and the extension cable next to the socket; and rewire so you end up with the flashing plug on one end and the extension socket on the other; then plug the wireless thingy into the socket. Stuff the lot inside the pony and attach a second extension cable to the exposed plug. That way someone else has already taken care of the worst of the soldering and data line power drain problems for you. Remember that whatever you ultimately plug it into will need to be able to supply power to the entire contraption. - MoonShadow



Apache again.  Virtual hosts this time.
I've got them almost working - I can see one with with one prefix and another with another prefix.
But how to I tell it "Divert any site I didn't specifically tell you about to *here*"?  It keeps randomly deciding from amongst the available sites, and if I tell it "redirect * to..." then that grabs everything.  --Vitenka

toothycat.net does precisely this. Put all the settings for the one you want to be the default outside of any virtualhost blocks, and make sure the virtualhost blocks override everything you want overridden for each specific host. I have one completely empty <VirtualHost *></VirtualHost> in my httpd.conf but don't remember if I need that. Oh, I also appear to have the line "NameVirtualHost? *" before any virtualhost definitions, don't remember what that does. - MoonShadow


Does anyone have a device for connecting IDE hard drives to a machine with no IDE headers which I could borrow (donor would need to see a Churchill before Saturday to effect handover)? The drives from my old machine have some data I want, but my new machine only has SATA controllers. I'd settle for an drive-IDE->mobo-SATA converter (I think I used to own one, or one in the other direction, so I'm pretty sure they exist) or some kind of external caddy (perhaps drive-IDE->USB). --CH


Does anyone have any experience with a data recovery organisation? I have a dead hard drive that has some non-backed-up data on it. Yellow pages lists three such firms that cover this area ([1], [2], and [3]), but I'm not sure which of them might be any good. It's likely that it is the drive electronics that are broken, but the data should be intact. On a related note, does anyone know of a fairly easy way to regularly back up ~1TB of data? --Admiral
A second hard drive?  --Vitenka (Yes, from the sounds of it, that's an eminently recoverable disk, but I don't have any recommendations on who to do it)
It'll already be on RAID on my new system. I was thinking a little more offsite. Regarding the hard drive, it might even be the computer (or the power supply) that has failed, rather than the drive itself. I'm going to try it in another computer at work to see if that's the case. --Admiral
As far as the second question goes, [backup2l] is the most sensible way of making backups of large volumes that I've encountered, but no idea if it scales to terabytes. - MoonShadow
Got my hard drive back from the data recovery company. I thoroughly disrecommend Rapid Track Data. They have made a complete mess of the whole thing, and were incredibly slow. --Admiral
What did they do (or not do) ? --RobHu
This calls for a list. --Admiral
Is that enough? Basically, their advanced data recovery software couldn't hack ext3.
Wow, that's quite a list of stupid things they did. As others have said below, wasn't http://www.forensicswiki.org/wiki/Dd_rescue dd_rescue] (or [ddrescue] which is superior) enough for recovering an image of the drive? I doubt their advanced forensic software is actually any better than the GNU stuff that's out there that I use when repairing drives for friends. --RobHu
My original notion (since the machine didn't boot up) was that the hard drive had a hardware failure. It was only when I bought a new monitor that was capable of switching on while displaying text mode that I found that Linux (after complaining for ages) was sitting there saying "Enter root password for maintenance or press CTRL-D to continue booting". Obviously, I couldn't repair a hardware failure, so I sent it away. Amusingly the bad sector number 81 (page two of the first partition) for some reason made Linux fail to recognise the partition table, after spending ages trying to read that sector. --Admiral

In the end, I hooked up the drive to my computer again. On boot, Linux complained quite heavily about a few blocks, but eventually booted up after failing to establish the partition table. I used dd to make a complete copy of the whole hard drive, marking bad blocks as I went. Note that by default when dd encounters a bad block it will write no data to the output file, so any further blocks will be shifted back by one - you need to fix that.
For future record, dd_rescue is specifically designed for this operation. --MoonShadow
There's an option for that - conv=noerror,sync should do it - bad blocks are then replaced with blocks of zero. --Vitenka
dd_rescue would have saved me a fair bit of time. If I had used dd conv=noerror,sync then to not lose any data I would have had to use iflag=direct and bs=512, which reduces the performance to 3MB/s. --Admiral
I found that there were 26 bad blocks. The first was block 81, which corresponds to page 2 on the partition, which is part of the superblock. That's why the system didn't like it very much. Using debugfs, I told it to read a different superblock, to analyse the rest of the bad blocks. The second bad block was in the middle of the journal, which I ignored. After that, there was a group of 21 bad blocks spaced about 1.6MB apart, which is obviously the cylinder size on this hard drive. Luckily those were all on the same file, which was itself a backup file and didn't matter. The remaining three bad blocks were on unallocated pages. So, I lost no data at all.

It takes a long time to copy 500GB of data. Heck, it took about ten minutes to delete the 500GB file afterwards! My new disc arrangement is 2TB RAID-1 (although I originally wanted RAID-5 of three discs, my motherboard doesn't have enough SATA ports).


CategoryComputing

ec2-34-205-246-61.compute-1.amazonaws.com | ToothyWiki | RecentChanges | Login | Webcomic
Edit this page | View other revisions | Recently used referrers | List subpages
Last edited July 28, 2010 11:43 am (viewing revision 37, which is the newest) (diff)
Search: