Conversation

Charlotte 🦝 therian

facts about lvm: this linux specific program that is built on top of linux specific features is well used on UNIX (which has no support for linux specific features like device mappers)

1
0
2

how to create symlinks on windows: use this feature which is in fact not a symlink and does not behave like a symlink

2
0
2

did you know that this /usr direction means “utilities and system resources” and not “user”? i mean that is only where you install all of your programs to and traditionally had your home directory in!

2
1
6

standard users on every system:

  • root
  • user that does not exist on my pc (6x)
  • nobody
2
0
2

also it claims that nobody has a uid of -2 which might be correct on systems with 16 bit UIDs, but systems use 32 bit uids now!

1
0
2

ah yes linux runlevels. for the 2 linux distributions that still use them

1
0
2

idk why the prof decided to expand what dig stands for, especially since it is not even an acronym anymore

1
0
3

ah yes inetd. the most relevant server application of them all

1
0
2

[about mail security] “nobody knows where the message is from” received from mailserver.test (192.0.2.7), received from firstname-pc (203.0.113.72)

1
0
2

including a warning that only applies to legacy version of html

1
0
3

@charlotte that’s one of the ultimate reasons why I have always tried to stick around *NIX systems. I mean, even macOS has symlinks.

that and windows weird habit of making a folder hidden by using a flag instead of a period at the start

0
0
0
Oh i forgot one: let's describe how inodes work based on legacy file systems like UFS, ext2, and ext3, not how inodes work in current file systems that are actively deployed, like ext4, xfs, hfs+, btrfs, and i think also apfs, because literally no modern file system aside from zfs uses blocks anymore, and zfs doesn't use inodes either
3
0
2
(modern file systems tend to be extent based, so instead of having 12 direct blocks, an indirect block, a doubly indirect block, and a triply indirect block, there is either a list or a B+ tree of pointer-length pairs)
0
0
2
@charlotte Fun fact UFS is still deployed since in some cases it can do some things more efficiently than ZFS, and it has lower memory requirements
0
0
0

@charlotte it's probably a retcon similar to b, i and u elements in html

0
0
0

@charlotte For my school I got this actually, I had to implement a FS driver for a UNIXv7 FS. It was something that actually was very useful to do, it thought us about low level file operations, reading RFC's etc. It's a bit of a missed skill with a lot of younger software engineers, not being able to do any low level file/binary operations

1
1
0

@maruno the context wasn’t even really about how filesystems on linux work, it was more “each file has an identifier called an inode” and then the prof pulls out implementation details from file systems that aren’t even used on linux anymore

1
0
1

@maruno like just the first part of the explanation would be fine as this

  • applies to all filesystems on unix, not just the ones with inodes
  • and is actually useful to know because if two files on the same device have the same inode, they are literally the same
1
0
1

@charlotte yeah, knowing the basics about inodes and file descriptors is important to know to understand low level file API

0
1
0