Ok ive been reading some stuff on udev and devfs and im a bit confused
as to why the static dev was even replaced. Im suspecting it was to make
device management eazier for novice users. in my own opionion udev and
devfs fail miserably at this. Both are hard to understand or get some things
to work properly.


Take for instance my lastest problem with udev. i put in a new harddrive
my existing parttions were in 3 in number with free space.

ok so i created a new one /dev/hda5 acording to cfdisk

then i try and mount
# mount /dev/hda5 /temp
< device doesnt exisit>

hmmmm so i wait but udev continues to not add the device node ?


so i reboot and yes it makes the node.


no biggie but why are we going to a windows 98 mentality that we need
to reboot the computer to be able to use a new parttion ?? This is stupid
thinking that windows 2000 or xp does not require this and i think linux
shouldent ither. Now there may be a way to do this but from a novice user
none of them are really dooable.


so again i say why create all these scripts and have this complicated garbage
just to use devices..... What was wrong with the static dev that needed fixed?

I agree udev (devfs is old and broken) is a bit confusing (read : real pain ). I was quite satisfied with the old mknod /mkdev and a big bunch of unrelated devices. It was quite alright as long you were not "missing" a device, in that case, it was a real nightmare to figure out what was the minor and major number of the device, if it was a block or a character device, etc... In that way, udev is quite simple for newbies, it's really easier to create a device now.
I believe udev was created in an attempt of standardisation. With all the new devices (USB drive, firewire, flash card, smart card, bluethoot, etc etc...) that suddently appear on market something like 5 years ago, it was a bit confusing to know what device was needed or not (what default devices you were needing on a new system). I believe they just decided to create some kind of "dynamic devices deamon" to create (and keep tracks of these) devices. This is indeed a really more mature way to do so anyways.

About your problem now... well, we are missing some technical details, but when you create a new partition on a drive, it is common that you need to reboot to reload the partition table (that would make udev to create the device as well). However, it is possible to reload "on the fly" a partition table using hdparm ( hdparm -z /dev/hdXY would do it ). Or sometime udev just forget to create a symlink or that kind of trick, you shouldn't need to reboot.

I'm not sure about hdparm, but I'm pretty certain that after you create a new partition and fs, you need to reboot in order for it to become visible and mountable.

Maybe hdparm eliminates the need for this?

As for udev, I think Half_Elf is right: when a device is lost, you're looking for a needle in a haystack, because the default /dev listing is HUGE. good luck looking for a USB mouse that you just plugged in and X is not recognizing.....

Yea i do use udev on my own system but i was just a little curious on why it dident create the device
but no worries.

static /dev was replaced with udev because static /dev is not adequate in terms of numbers and types of devices. In order for the system to be truely scalable for any potential aplication and any conceivable future devices /dev must be dynamic.