| joerg | firmware? | 00:20 |
|---|---|---|
| plasma41 | archdummy: I've cloned both of those repositories to my local system. I'll take a look at them whenever I find time to set up snapper on my machine. | 00:34 |
| integfred | Thunar file manager doesn't copy files using the GUI to external media properly. Copying by GUI results in file corruption. You can test this with cmp comman | 01:00 |
| integfred | Using a different file manager (spacefm-gtk3) works for the time being | 01:00 |
| integfred | It's important as I just ripped a computer apart trying to find the source of corruption | 01:01 |
| integfred | Pretty severe | 01:01 |
| ted-ious | integfred: What filesystem are you using? | 01:07 |
| integfred | ext4 | 01:07 |
| ted-ious | On both? | 01:07 |
| integfred | yes, I used a 6GB file to test | 01:08 |
| integfred | copying by CLI was fine and copying using spacefm was fine | 01:08 |
| integfred | but NOT Thunar | 01:08 |
| ted-ious | What kind of file? | 01:08 |
| integfred | I tried an ISO originally, then I dded some random stuff to a file to try that as well | 01:09 |
| integfred | It only happens with external media | 01:09 |
| ted-ious | And you're testing for corruption with something like sha*sum? | 01:09 |
| integfred | cmp | 01:09 |
| ted-ious | Oh right. | 01:09 |
| integfred | I'm testing one file against the other with cmp | 01:09 |
| integfred | but shasum would work too | 01:10 |
| ted-ious | Ok let's see if we can devise a test that will determine the nature of the corruption. | 01:10 |
| integfred | I also mentioned this on xfce IRC just now | 01:10 |
| integfred | dd if=/dev/urandom of=somefile bs=1M count=6000 | 01:11 |
| integfred | copy the file via gui | 01:11 |
| integfred | then | 01:11 |
| integfred | cmp somefile /path/to/externalmedia/somefile | 01:11 |
| ted-ious | Does it corrupt with just one small file? | 01:11 |
| integfred | I haven't tried one small file | 01:12 |
| integfred | but it's possible | 01:12 |
| integfred | I was getting corruptions because I wrote a script to check for silent corruption | 01:12 |
| ted-ious | I would try making a simple script to do larger and larger files until the cmp fails. | 01:12 |
| ted-ious | That's good. | 01:12 |
| integfred | https://github.com/publicsite/consistencyChecker | 01:13 |
| ted-ious | I started using archive file formats for copying files around for the same reason. | 01:13 |
| ted-ious | Or else creating a md5sum file and comparing it on the other end. | 01:13 |
| ted-ious | 7z and tarlz are really good for this. | 01:14 |
| integfred | i'm in the process of trying to make it use CRC-24 sum, converted to base64 (which is 4 bytes) and using a 4 byte CRC per 4KB | 01:14 |
| ted-ious | You could just use tarlz and it will do all the checksums internally for you. | 01:15 |
| integfred | yeah, i don't want the compression | 01:15 |
| ted-ious | That just needs a -0. :) | 01:15 |
| ted-ious | Although you can get minor compression for free with -3 since your cpu will be bored waiting for the usb and wants to do something more than twiddle its thumbs. :) | 01:16 |
| integfred | yeah ... I have slight paranoia haha | 01:18 |
| ted-ious | Trying to avoid any possible corruption? | 01:19 |
| ted-ious | I respect such paranoia. :) | 01:19 |
| ted-ious | You should read about tarlz then. | 01:19 |
| ted-ious | I think it's the only foss archive format with compression that can both detect corrupt bits and recover from bad blocks. | 01:21 |
| integfred | actually i was part way through writing recovery too | 01:21 |
| integfred | but this is not in the git yet | 01:21 |
| ted-ious | Have you looked at par2 yet? | 01:21 |
| ted-ious | There's supposed to be a par3 in development but it will take years before everybody is satisfied and trusts it as much as par2. | 01:22 |
| integfred | I have never tried par2, I guess what I'm after is similar | 01:23 |
| integfred | but part of the fun for me is to try and reinvent the wheel :) | 01:24 |
| ted-ious | Don't let me stop you then. :) | 01:24 |
| ted-ious | Does your script do things that rsync -c doesn't? | 01:27 |
| integfred | i don't know what rsync -c does | 01:37 |
| ted-ious | It forces rsync to compare every file instead of just looking at the date and time. | 01:54 |
| integfred | the script I wrote in the git at the moment, as it stands, basically does that (compare every file in 2 different directories against each other). I am working on the other things I mentioned like CRC check and recovery to replace the cmp aspect to clarify. I am doing quite a bit of the work in C, but I haven't uploaded that bit yet. | 01:59 |
| integfred | It's also worth noting that the bug in Thunar is distributed with devuan iso if you didn't already know this :))) | 02:47 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!