Distribution Watch: Gentoo Linux 1.2: Getting Better Every Day - page 3
Back in the Saddle with Gentoo
Gentoo offers a nice set of installation instructions. Once I downloaded the smaller ISO and burned the CD-ROM, I popped the media into my lab machine and booted the box. The language selection and PCI auto-detection as usual passed by smoothly, and in under a minute I'm at the base installation console.
The first instructions are to load the appropriate networking card modules, but instead I type lsmod and see that mine was loaded after being properly auto-detected, so I don't bother. Running dhcpcd eth0 gives the expected response with errors, though different than the errors I got the last time I went through this process. Still, a quick ifconfig tells me that I'm still doing fine.
Following the handy dandy partitioning chart they give in their installation documents, I utilize fdisk to create the recommended set of three partitions (/boot, swap, and root), and change the swap partition's type to 82. From there it's mkswap to make the appropriate filesystem on the swap partition, and then mke2fs j to create a nice robust journaling ext3 filesystem on the others.
Now to mount the partitions. A quick swapon for the swap partition, and then I forget that I do literally have to follow the order given while creating the mount points and adding the partitions, but that's an easy thing to fix. Then soon I've got the Gentoo CD-ROM mounted on the CD-ROM drive, and visible for the filesystem. This is where I hit my first snag, but it was from looking at too much code at once, or sunspots, I'm not sure. Once I properly changed to /mnt/gentoo and then extracted the Stage 1 tarball to the proper directory, everything worked fine.
Once I've mounted the proc filesystem, copied /etc/resolv.conf over, and chrooted to /mnt/gentoo, I cross my fingers. If I remember correctly, here is where I had problems last time. Oops, there we go again. Typing env-update gives me an error, but it also gives me the output that the installation guide tells me to look for (after the error). So, I type source /etc/profile and lo and behold, there's a prompt! Considering that I never was able to get that prompt last time, things are looking promising.
Now for the first download. I type emerge rsync and Portage takes over my cable modem, downloading 10MB worth of material. Last time I ran this I was using ADSL but my connection was intermittent, so I had to download all of this material half the time under ADSL and half the time under a regular 56.6 kpbs modem. From that experience, I'd say that if you don't have a high-speed connection, you're probably better off getting the larger CD image and saving yourself some immense frustration. Maybe not so much now, but later in the process.
Since I'm using the Stage 1 only tarball, I now get to bootstrap my machine, but first I have to tell Portage how to handle the compilations. Instinctively I try to utilize vi but the text editor included here is nano (pico, nano, get it?) So, nano w /etc/make.conf and I select the P6-exclusive settings so I can fully optimize the OS, and then save and exit. And now it's bootstrap time. I change to the Portage directory and run the bootstrap shell script, and then go off to do something else. After all, the installation instructions point out that on a 900MHz AMD Athlon this process can take an hour, and my poor lab machine is only a 450MHz.
However, I come back later to find that the bootstrap has stalled with an error. Being the odd superstitious sort that I am, I try it a couple more times, and nothing changes. At least it's consistent. I go dig through the Gentoo Bugzilla entries, the discussion and troubleshooting forums, and find a few things that might be relative but nothing close enough. Rather than immediately filing a bug I figured I'd better try the sequence again to find out if it was repeatable, and if perhaps it was just a glitch.
This is where things went bad, and part of it isn't any fault of Gentoo's unless the folks at Gentoo Linux have the ability to cause trouble at a nearby power substation. First, I think the bootstrap managed to finish but the machine hung overnight with the same problem I'd run into earliera blank screen as though the terminal had gone into wait mode but it couldn't recover. So I rebooted and assumed the bootstrap had worked, and went through the parts of the process that load the partitions back into place and then proceeded from there.
Meanwhile, during the day, we lost power at least three times. Since we just moved and have rearranged all of our machines I'm in the annoying situation of having a single UPS that can only handle one computer and one monitor, and my lab machine is not the protected one. Let's just say that when I tried proceeding to a later stage and thing s went wrong again, I decided it was time to do everything over on another machine. Between the power outages and the previous version never learning to get along with the other hardware it seemed a good idea.
So, I made good use of the idle hard drive in my main Linux box. Everything worked like a charm. Got through the bootstrap without a single difficulty. Passed from stage 2 to 3 with no problem. I finally get to set the local timezone, and I select the vanilla-sources version of the kernel since I'm not trying to do anything wild and crazy and I feel like being a little different than Gentoo's favorite kernel (plus, let's see how it works with a more vanilla kernel).
In configuring the kernel, the first thing I do is address the issues laid out in the warning in the Gentoo documentation. I'm a bit forgetful so I might miss them otherwise. I tweak a few things here and there but for the most part I'm not looking to do anything too out of the ordinary. I don't have a compiled kernel yet so I have no bzImage to move. I then select syslog-ng since I've been reading about that logger lately, and I select vcron since that's the one I'm familiar with.
Now I have the fun of setting up /etc/fstab (this part certainly could be better documented for folks newer to Linux), including setting the proper devices to the proper partitions, my filesystems to ext3, and my options to defaults. Then I set the root password, create /etc/hostname and tell it what to call my machine, leave /etc/hosts alone because the only machines this one will deal with are all resolvable with the information given to my Linux box by DHCP, configure the networking, edit startup files, and then finally set up GRUB.
Since I'm setting up a dual boot box, I actually can't complete the GRUB setup until I reboot and go into Red Hat Linux, which has the primary hard drive and already has GRUB installed. Once I get Gentoo added to the Red Hat installation's /boot/grub/grub.conf I'm ready to go!
Solid state disks (SSDs) made a splash in consumer technology, and now the technology has its eyes on the enterprise storage market. Download this eBook to see what SSDs can do for your infrastructure and review the pros and cons of this potentially game-changing storage technology.
- 1Linux Top 3: CoreOS, Oracle Enterprise Linux 7 and Ubuntu 14.10
- 2Linux Top 3: Debian Dumps SPARC, Ubuntu Takes Over Linux 3.13 and the Core Infrastructure Initiative
- 3Linux Top 3: Fedora, Ubuntu and Gluster Lose Community Leaders
- 4Red Hat Enterprise Linux 7 Finally Hits the Big Time
- 5Linux Top 3: Tails 1.0, OpenMandriva Lx 2014.0 and Debian 7.5