The old maxtor's Karachun came on an old server with Gentoo and the prehistoric date of the last launch of emerge -uDpv world. Somehow, I didn't want to install-compile again. Blown by it in due time. I wanted stability, reliability, speed of installation of updates (the server is far from the first freshness) and the absence of problems with dependencies. In general, laziness is the engine of optimization.
At first I wanted to stick FreeBSD and do not bathe, because I stuck it, it works, I forgot where the server is. But the server is not simple, but for developers, more precisely, various Linux binaries (no source) are running around it — it is not compiled, and the architecture is -32bit. Hence the limitations - Linux distribution, 32 bits, with wide support for old hardware, availability of old libraries, new technologies, graphics disappears, ReiserFs out of the box, because Separately there is a screw with a partition / home and with this same ReiserFS (3.6, by the way). From services - Samba, DNS, DHCP, TFTP.
How it was
Modern trends suggest a wide choice of Linux-distributions for the role of the server operating system, and the first of them was tested by Ubuntu Server 12.04. Installation - excellent, smart, updates installed. The partition with / home, after writing the required line in / etc / fstab and rebooting, did not take off ... It ran through the logs on the screen that there is no such partition on your disk (there is one). On the second disk (also 1 partition with ReiserFS) is there, and it was successfully mounted. And on where / home - no. Yes, the disks are naturally tested by MHDD and fsck.reiserfs. Moreover, fdisk -l also shows that there is no partition, but if you open a disk through fdisk, then there is a partition, I write the partition table (the w command), exit is all, the partition is visible and mounted. Rebooting - there is no partition, fdisk, there is a partition ... Yes, he himself is in shock. Ubuntu Server is not suitable.
So, returning to the topic of the story, CentOS was chosen, due to the experience of server administration under its management and the experience of resolving the jambs under its management.
After installing the fresh image of CentOS 6.4 (I’ll skip the details), it turned out that ReiserFS is alien to it in the default installation. Good Google has carefully issued several links to English-language resources on this topic, which, in turn, have links to other English-language resources on this topic ...
')
Short
course of the young fighter list of actions to achieve the goal:
We need to install another kernel — the centOS-plus kernel and the reiserfs-utils toolkit.
But first the core. We include CentOS-Plus repository. Information about this in bulk, therefore in brief:
Rule /etc/yum.repos.d/CentOS-Base.repo, block:
[centosplus] ...... enabled=1 ........ includepkgs=kernel* jfsutils reiserfs-utils
Update the kernel:
yum install kernel
Should write something like:
….. kernel-2.6.32-358.6.1.el6.centos.plus.i686 …...
Yeah, right core. You can check in grub.conf (menu.lst) just in case if it is registered as default when loading.
The utilities of reiserfs-utils are in standard repositories with the command:
yum search reiser
missing:
…. Error: Nothing to do
As it turned out, the necessary reiserfs-utils utilities are in the
ELRepo repository. We connect it.
Import the public key:
rpm --import http://elrepo.org/RPM-GPG-KEY-elrepo.org
And the repository itself:
rpm -Uvh http://elrepo.org/elrepo-release-6-5.el6.elrepo.noarch.rpm
Now:
yum search reiser
gives what you need:
…. ===== N/S Matched: reiser ===== kmod-reiserfs.i686 : reiserfs kernel module(s) reiserfs-utils.i686 : Tools for creating, repairing and debugging ReiserFS filesystems.
We put:
yum install reiserfs-utils
I rule / etc / fstab to fit my needs, reload:
I am pleased.
ps eyeglasses - please, in a personal, correct. Inaccuracies, comments, questions - in the comments I will gladly accept / answer.