Steps after installing GNU / Hurd

[Monty1]

Once GNU / Hurd is installed, you will start immediately. But unfortunately you have to set up and configure a few things before. Here are some useful tips on how to rule to the first cliff.

Network setup

Setting up a network, ie the integration of the Hurd-computer as a LAN, is fortunately quite simple. GNU Mach must recognize the network card, otherwise you're nothing. If you have when you do not cross-install specified a separate host name, the computer will recognize as "hurd". In retrospect, this may be the altering of the entries in / etc / hosts and / etc / hostname of "hurd" to "Scotty" to do such. The next step is to "connect" to the corresponding socket with the network card. This is done by calling:

settrans-fgap / servers/socket/2 / hurd / pfinet - interface = device \
- Address = nn.nn.nn.nn - gateway = GG.GG.GG.GG - netmask = MM.MM.MM.MM

The options "fgap" cause a previously furnished Translator is removed and a new passive translator is constantly added. / Servers/socket/2 is called the inode of the Translator / pfinet is linked / hurd. The other options are then to the translator. With --interface network card is the selected (the first detected network card can be reached via eth0, the second on eth1, etc.) --address is the IP address of the Hurd-computer --gateway is the IP of the gateway and --netmask is the netmask. The values are given in the usual format, for example --interface=192.168.1.1 . That's it. If a call is from ping is successful, you've made it. If not, here's a hint: The server / hurd / pfinet is a no typo error message. Incidentally, this concept plays a central role of translators in Hurd.

[Monty1]

Mount filesystems

The concept of mounting file systems is compared to GNU / Linux very different. As before, it works again via a call to settrans. In this case, a different translator (server) is required. Currently only ext2 and NFS formats are supported. Also applies to ext2 partitions, the restriction that it is not possible to mount a file system that is larger than 1 GB. This restriction does not apply to NFS. You can mount an NFS volume as follows:

settrans-fg / mnt / hurd / nfs 192.168.1.2: /

After that the directory / mnt connected with the root file system of the computer with the IP address 192.168.1.2. The NFS server provides some useful options. A direct call to /hurd/nfs --help it shows them. This is also true for other servers. The mounting of ext2 file systems is not directly, because most likely still a few device files. These are created as follows:

cd / dev
. / MAKEDEV hd0s2

"Hd0s2" stands for the second partition of the first IDE hard disk. can be device files for SCSI disks to create one as follows:

cd / dev
. / MAKEDEV sd0s3

Unlike GRUB IDE and SCSI drives can be distinguished. Hard Linux users will have noticed it: hd0s1 or sd0s1 is the same as in GNU / Linux hda1 or sda1. Here is a small example, just to be safe: a system with two IDE HDs and a SCSI HD would look like this:

• hd0s? <- The first IDE drive (master on the Primary Channel)
• hd1s? <- The second IDE drive (slave on the primary channel or master on the secondary channel)
• sd0s? <- The first SCSI disk

The "?" stand for the partitions. The first partition is numbered "1" and not "0". The partitions 1-4 are primary partitions, one window may also be (extended) extended. This makes it possible to create more than four partitions. The extended partition contains at the logical drives. They will get from point 5 This applies to both IDE and SCSI drives. After then, the device files, you need to connect only the ext2fs server with a "node". This goes as follows:

settrans-fg / mnt / hurd/ext2fs / dev/hd0s5

Again, you can find detailed information about possible options, by a direct call to /hurd/ext2fs --help conduct business. The settings are the way permanently, ie even after boot, the partitions automatically be included in the directory tree. If this is not desired, you should remove before the computer shuts down the translators:

settrans-fg / mnt

[Monty1]

Swap partition

RAM you can never have enough, but since this can be an expensive undertaking, is usually not enough memory in the computer are available to meet Hurd. Especially if you plan to translate the libc0.2 (100 MB RAM and 200 MB swap required plus 800 MB on the hard drive!). Therefore you should set up a sufficiently large swap partition. Use fdisk to set up a partition of type Linux swap (82) field. It is no more limited than some Linux kernel series 2.0 and 2.2. If you already have swap partition is one that is used by Linux, can be an entry in the file / etc / fstab to make simple:

# <file System> <mount-point> <type> <options>
<dump> <pass>
/ Dev/hd0s6 none swap sw 0 0

Of course there must be the associated device file:

cd / dev
. / MAKEDEV hd0s6

Then you only have to swapon /dev/hd0s6 activate the swap partition.