[–] 8 points 6 months ago (6 children)

Generally, at least in somewhat decently set up systems, attacker has no way to know if the password was 4 or 40 characters long. Also, in general, attacker can't just try every possible password, as it would just take too long. With short passwords (and limited set of characters) it's possible, but with tens of characters it's just not going to happen.

Additionally, if the system is built even half competently, passwords are stored in a database in hashes and with a salt, so if your password is 'hunter2' the value in database would be something like 'supersecretsalt+hunter2' (trough hash-function obviously), so you can't just take rainbow tables and compare them directly.

Now, if the attacker has the salt then they can build their own tables with known passwords and see what matches or in some cases they can reverse engineer the salt out of the system and proceed with "normal" approach. Or, depending on the breach, they might have access to the password verification functions and they can just override rate limits and other stuff and brute force accounts locally and then exploit them at the real world.

There's a ton of nuances on this and I'm not an expert by any stretch, but in general if you use sufficiently long passwords (preferably with a password manager) there's no way to brute force your password with current technology. But, since there's 'a few' breached passwords around the net, it means in practise that not all systems are built securely and their vulnerabilities are used to get account information, passwords included. Also malware, phishing and other means are possible when gathering login info.

With a new data breach attackers might just be happy that they can break 20% (or whatever) of the passwords and use them as they see fit and just ignore the rest as it's not really worth the time to try and break them all. And if they are interested in some spesific account then there's other ways, like phishing or social engineering, to build attacks targeted on that spesific individual.

  • source
  • [–] [S] 1 point 6 months ago* (1 child)

    The reason that I didn't think that you are right is that I've literally copy pasted the credentials from the ISP portal.

    Level 1 support, when he asked my credentials - did not raise any issue when I repeated them to him.

    I did end up talking to a lvl 2 support.

    He did a reverse-facebook-pitch "drop the the" and by the he meant the ".net" suffix at the end of the username - AS IT IS SPECIFIED ON THE ISP PORTAL.

    This was a week long frustration that was never meant to be.

  • source
  • parent
  • context
  • [–] 2 points 6 months ago (2 children)

    Says right in the logs your PAP auth is failing, so you have something off. Could be your auth method, could be your credentials, or could even be it's being rejected because your SD-MAC changed (did you compare them?). You need the logs from the other side to be able to tell.

    What I would do is contact your ISP and ask for Level 2 tech if some sort and get the business/enterprise config for PPPoE devices. Tell them you need the manual configuration for PPPoE for your device, and just work through those settings. They also be able to check logs and see which step your auth is failing at.

  • source
  • [–] [S] 1 point 6 months ago (2 children)

    It wasn't a ISP configured router. It was entirely mine. I do suspect different things being sent in stock/openwrt. If I was able to revert to tplink firmware I would chase this lead, but for now - I bought another TP link router, waiting for it to arrive, then arranging a technician to pair the new router with the GPON (thus abandoning OpenWRT efforts entirely)

  • source
  • parent
  • context
  • [–] 1 point 6 months ago* (3 children)

    I haven't used PPPoE myself, but if you can't get it working and you think that you're sending the same stuff, maybe get yourself another ISP-configured TP-link ("I broke mine and need a new one" and I imagine that they send one out), then either use the first device or another Linux machine with at least two Ethernet ports between it and the ISP, bridge the two ports, and use tcpdump to capture a pcap file containing the traffic and open it up in Wireshark to see what the two are sending differently?

  • source
  • [–] 2 points 7 months ago

    HDDs have plenty failure modes. You might get SMART warnings. Or maybe you don't and fails a bit more spontaneously. I had harddisks do that nasty clicking sound because something's wrong with the heads. And sometimes I could still read data from the other areas of the disk. Sometimes it would refuse to read anything. And sometimes a HDD just works and the next moment it won't spin up again, or the controller is dead.

    I think the only reliable method is backups.

  • source
  • parent
  • context
  • [–] [S] 2 points 7 months ago (1 child)

    Thanks, that gives me some options at least, even though there's little hope. It went from plenty fast and healthy two days ago to basically dead yesterday, without much warning. At least as far as I could tell.

    I see these articles talk about "failing", but what state would that have been? Should I have seen the signs and prepared? (Though I think "prepare" would've been the standard advice of backups, second drive, second location.)

  • source
  • parent
  • context
  • [–] [S] 1 point 7 months ago

    It is an HDD and it's about 10 years old, so I'm assuming the worst. I don't know much about failure modes of hard drives, whether it would be mechanical, electrical, magnetism gone stale...

    Anyway, I'll disconnect it for now, order a fresh drive and try to ddrescue the image to that. Seems doable at least.

  • source
  • parent
  • context
  • [–] 2 points 7 months ago* (2 children)

    The Arch Wiki has a comprehensive article on Data rescue: https://wiki.archlinux.org/title/File_recovery

    And there's more in the Ubuntu Wiki: https://help.ubuntu.com/community/DataRecovery

    What I do is run ddrescue and scrape all I can get from the disk. Into an image file on another disk. Then I'll unplug the broken disk. If that's enough data, I'll run testdisk to recreate the partition tables, extract files... And if that fails, we have software like photorec.

    Not sure if ddrescue is able to do anything in your case. Sometimes a harddisk will just fail entirely and there's nothing to be done, except swap controllers or other broken components in a clean room.

  • source
  • [–] 5 points 7 months ago (1 child)

    Is it a HDD or an SSD/NVME?

    It's usually a very bad sign if all of a sudden you can't see partitions anymore on the drive.

    Basically, if you manage to boot and the drive gets properly recognized, you can either try to copy stuff off of it using standard methods. Or you can try some more advanced tools, e.g. ddrescue.

    Sometimes you can get lucky, and it's just a loose connection.

  • source
  • [–] 3 points 7 months ago (1 child)

    In general shader caches are compiled for the GPU hardware and driver they intend to run on, which means you can't precompile all of them if you don't know what hardware setup the software is going to run on. This isn't unique to emulators either. PC games have this same problem and either stream it in the background or hide it in loading screens.

    The difference for emulators is that they don't have the benfit of source code to work from and effectively have to decompile, translate, and recompile. These steps can change each version as translation accuracy improves but ultimately means starting over with a blank cache. For simpler consoles before unified shader pipelines, this isn't much of an issue because there are a fixed number of render methods.

    In summary, you can really only share caches if the GPU, driver, and version are all the same. Otherwise the emulator will invalidate and recompile the cache.

  • source
  • [–] 1 point 7 months ago

    Booting to the pre-install environment was the goal to finger obvious hardware/stability issues. Something that lands you on a desktop. If you want to try installing - whatever distro it was - check a guide for it. Fedora and Ubuntu are good places to start for both of these routes. They should require minimal configuration if you don't care about secureboot etc.

  • source
  • parent
  • context
  • [–] 1 point 7 months ago*

    I have a really hard time believing that both the Windows installer and a Linux installer don't work without there being at least one hardware problem present. But you did swap out most of the parts that I'd think are likely culprits.

    Hmm.

    When you were swapping parts on this thing, did you have the system either unplugged (preferably) or powered down at the PSU switch? Not just the motherboard's power off? I remember once accidentally pulling a video card without cutting the actual PSU power, frying both the motherboard and (though I didn't realize it at the time) the video card, then plugging the video card (correctly) into a new motherboard and frying that motherboard as well, so that even after I swapped out the video card, I still had problems. That is, a damaged part theoretically can damage other parts. Never heard of anyone else hitting something like that, but...

    I could imagine that an underpowered PSU could maybe cause a variety of other failures, but you said that you replaced it, and I'm assuming that you checked that it was rated for the components that you had.

    Manually underclock the memory in the BIOS as far as possible? Try running with just two sticks (your motherboard manual will tell you which slots to use if you have only two DIMMs)? If the memory's marginal with your hardware, that will help. Not saying that that's a fix, but it'd isolate the issue.

    EDIT: Maybe vacuum the thing? I guess if you have metal shavings or a screw or something like that rolling around in your case shorting stuff, it could cause issues.

  • source
  • parent
  • context
  • [–] 1 point 7 months ago* (last edited 7 months ago)

    I mean, something clearly isn't working right there, and I can't think of many things that would cause that.

    considers

    The keyboard (and mouse?) do work in the BIOS, just not in the Windows installer? While plugged into the port in question? (I mean, it's possible to disconnect the cable running from the USB header on the motherboard to the front panel.)

    Have you tried, while the PC is on and the Windows installer is active, unplugging the keyboard and mouse and then re-plugging them back in? That'll reset the device and let the installer handle them. It's not completely impossible, I guess, that the BIOS talking to them could somehow put them in some kind of wonky state.

    EDIT: If the keyboard and mouse aren't working in the BIOS either, try them on someone else's computer to rule them out as a source of trouble? I doubt that they are the problem, but might as well get the data point, just to be sure.

  • source
  • parent
  • context
  • [–] [S] 1 point 7 months ago (1 child)

    Have you tried plugging them directly into the USB ports on the PC?

    ...yes. since the first attempt over a month ago. Front ports. Motherboard ports. And i never plug a mouse or keyboard into my monitor or other usb hub.

  • source
  • parent
  • context
  • [–] [S] 1 point 7 months ago (2 children)

    Can you boot to USB? Using something other than Windows may also help determine that the hardware is working. If everything seems to work, it could be an issue with the boot entries on your main board, corruption on the disk, or something wrong with the disk physically. If you use Linux, whatever the latest Fedora version is will probably support newer graphics cards such as a 9070xt.

    I still cannot, for the love of AntiChrist, boot the thing. At. All. Not even from a usb.

    I tried installing Linux yesterday. Just an eternal black screen for hours and hours after selecting "Install Linux".

  • source
  • parent
  • context
  • view more: ‹ prev next ›