PHP usergroup parser (/etc/group)

Either I'm too lazy to have searched for an existing method, or there isn't one.

So long as your system isn't confined by open_basedir or something like that, this should work. Personally, I need this to get the names of the groups by their numeric ID without having to use a bunch of calls to filegroup() or by using DirectoryIterator::getGroup, although now that I take a closer look at the latter it seems like a better and better idea. Originally, I was going to use stat() and parse the results from what my script finds in /etc/group.

I might still use this for something or other, or perhaps not. Either way, here it is for your using and viewing pleasure.

The script will read the contents of /etc/group and return an array where the keys are the user group names and the value is their numeric id. That's just how I whipped it up though it would probably make more sense to have it the other way around.

<?php

$g = file_get_contents('/etc/group');

$g = explode("\n", $g);
$c = count($g) - 1;

$groups = array();
$x = 0;
foreach ($g as $row) {
  if ($x === $c) {break;}
    $row = explode(':', $row);
    $groups[$row[0]] = intval($row[2]);
    $x++;
  }

echo json_encode($groups);
?>

Change MySQL root password

After initially installing MySQL on any platform your root user usually has no password. To set it from the command line issue the command:

# mysqladmin -u root password NEWPASSWORD


That's pretty basic stuff, but what if you get locked out of your server from having a typo in the password set, forgetting the password, etc, etc?

First, you need to stop the mysql daemon from running.
Then start the daemon again without the grant tables (the tables containing permissions and user data), and if you are on a live server, you may want to start it without networking since without the grant tables ANYONE can access ANY database from ANYWHERE.

Note: If you are running as root user on your system you will need to add the --user=USER argument to the following command, where user is the user that mysqld will run as (usually mysql)

# mysqld --skip-grant-tables --skip-networking [--user=mysql] &


Now you can log in to mysql without a password or user:

# mysql


Now you just need to update the root user password using mysql:

mysql> UPDATE mysql.user SET Password=PASSWORD('NewPass') WHERE User='root';


Then flush the privileges table:

mysql> FLUSH PRIVILEGES;


After that is done, you need to stop mysqld and restart it normally. You will now be able to gain access to the root user with the password you have set.

Installing Fedora 16 from Hard Drive

Okay!

To start, this post will assume that:
- You are using the grub boot loader. That's it.

For reference, I use different partitions. My disk is partitioned in this manner:

Device Mount Point Size Description
/dev/sda1 /boot 4 GB Used to hold kernel and fs images.
/dev/sda2 ~320GB Extended partition labeled canister to hold logical devices
/dev/sda3 ~130GB ext4 filesystem
/dev/sda4 / ~90 GB Partition that holds Arch Linux, which I use as a fallback.


The sizes are probably off, irrelevant, I know.

Now, go download the Fedora 16 ISO image. *NOT* the LiveCD image, the entire install DVD.

After you have this, mount it:

# mkdir -v /mnt/iso
# mount Fedora-16-i386-DVD.iso /mnt/iso
# cd /mnt/iso


I might mention here that if you have your /boot on a separate partition as I do, you will need to ensure that it is not very small. I have mine at 4GB, the default size was 100Mb so I had to grow the partition to fit the Fedora vmlinuz image, which is ~128MB alone. You can check by typing df -h
You may not need 4GB. I will leave it up to you to do the math. If you have your entire installation on one partition, including /boot then I guess you don't have to worry.

ANYWAYS, carrying on.

Being in the mounted directory, you will see a folder named `isolinux`, cd into this directory.
Be sure that you have your boot partition mounted and copy these two files into boot, ensuring that you rename them so you don't overwrite anything that might be there already. If you do that, I LAUGH AT YOU BECAUSE YOU CLEARLY DIDN'T READ WHAT I JUST TOOK THE TIME TO WRITE.
Let that be a lesson because you can't undo it and you've just overwritten your working OS with an installation image. If that's okay with you, just reboot and it should boot right into the install. If not, HAHAHA.

Do this:

# cd isolinux
# cp vmlinuz /boot/f16install-vmlinuz
# cp initrd.img /boot/f16install-initrd.img


Now people would say that "oh, you should edit your grub configuration!", but you know what? I find it complicating and annoying to have to add it into your grub menu. Just boot it from your grub command line, which you can access by pressing 'c' at the grub menu:


grub> kernel=/f16install-vmlinuz
grub> initrd=/f16install-initrd.img
grub> boot


The second command might take a little bit to load the image, but it will. Don't be so impatient!

That's it. Enjoy.
Any problems aside from "I didn't read the instructions and I baleted my working operating system, WHAT DO?" I will be more than glad to help.

GRUB shell on arch linux

My previous post mentioned how to boot from a rootfs shell. From what I am experiencing, I screwed up the grub configuration, so when I boot my computer I am entered directly into a grub shell. To get into the rootfs shell so I can boot arch linux ,I issue the following:


grub> kernel=/vmlinuz-linux [ENTER]
[Linux-bzImage, setup=0x4200, size=0x2fc90]
grub> initrd=/initramfs-linux.img [ENTER]
{Linux-initrd @ 0x7fd35000, 0x2ba04e bytes]
grub> boot [ENTER]


It then proceeds to boot the kernel. I am aware that it's taking me to rootfs because I'm not entering the device. I just haven't toyed with it enough, or bothered searching for the appropriate values. /dev/sda and /dev/sdaX do not work. I will update whenever I figure it out.

UPDATE:

I found out what was causing this. The entry in grub (which I have now agonizingly updated to grub2) was set to look for the device /dev/disk/by-uuid/c48c81b7-e772-42a6-95af-f27d768b887b
looking in /dev/disk/by-uuid, that uuid did not exist, nor was it the one for /dev/sda4
I changed it to the corresponding uuid and what it was linked to.

Here is a sample output from my laptop:

# ls -l /dev/disk/by-uuid

lrwxrwxrwx 1 root root 10 Mar 29 14:46 1c7fe2bc-2f14-48e4-84b7-08f1c53a90e2 -> ../../sdb1
lrwxrwxrwx 1 root root 10 Mar 29 14:46 492b94e8-25b4-463b-86a3-6e80d3b40632 -> ../../dm-2
lrwxrwxrwx 1 root root 10 Mar 29 14:46 4c34416c-fc03-4add-b1a0-ba9f5603bc4c -> ../../dm-0
lrwxrwxrwx 1 root root 10 Mar 29 14:46 78b772e4-59b1-4563-b6ab-6d9c24ef9fcd -> ../../sda2
lrwxrwxrwx 1 root root 10 Mar 29 14:46 c3230d94-1048-4c3e-8e35-aea175f6f667 -> ../../dm-1
lrwxrwxrwx 1 root root 10 Mar 29 14:46 f8c7dfa7-907a-4ac5-a2e4-a89e139b7224 -> ../../sdb2


Pretty straightforward, no?

service script on arch linux

Personally I am accustomed to redhat builds of Linux and find myself typing "service".
Instead of typing /etc/rc.d/$service $command where $service is the name of the service I wish to start, or searching for a pre-made solution, I wrote up a little script and saved it as /usr/sbin/service.


#/bin/sh
/etc/rc.d/$1 $2


After saving it chmod it to 755 so it can be executed. You can now issue the command: service network start

and have the network service start successfully, or whichever service script you find in /etc/rc.d

rootfs on arch linux

Find yourself in a bind and logged in to a recovery shell named rootfs?

If you happen to know the partition that your operating system is on, and the boot partition, you can boot into your kernel from this shell.

On my system, I have this layout:

/dev/sda1 > /boot
/dev/sda2 > /home
/dev/sda3 > / (ext4)
/dev/sda4 > / (ext4)


From this rootfs shell, I executed the following commands and booted into the desired kernel on sda4.


# mkdir /boot
# mount -t /dev/sda4 /new_root
# cd /new_root
# exec init


This had me boot into my fresh install of arch that I had installed from another partition. Now I can fix grub!

Rogers, Y U SO SLOW??

This is a post the is only very slightly related to anything I'm working on, but when I go to download things off my local machines to a remote machine in Chicago I am forced to upload a 1.2 KB/s or LESS!!!!!

I am paying for the Ultimate, read that again, Ultimate package. 50Mbps Downstream and some other 2Mbps Upstream. This is not torrents, it is yum package manager downloading remotely from my local filezilla server.

You're all liars and if it weren't for the extra hundreds of dollars that I'd rather not spend going to another company, I would do just that.

Installing Guest additions in VirtualBox; Windows 7 host.

In order to share folders, clipboard and other things you must install the Virtualbox Guest Additions. The iso file that you need to mount in your CentOS guest is located in the VirtualBox installation directory and is named VBoxGuestAdditions.iso.

EDIT: I just realized that the option to mount the guest additions iso file is right in the menu! From the guest window, click on Devices > CD/DVD Devices > Install Guest Additions
If you don't see it there, you are probably using an older version of Virtualbox and should follow the directions below.

From the guest window, click on Devices > CD/DVD Devices > Choose a virtual CD/DVD disc file...
You might see the file name in the list already, but if not, you need to navigate to the installation directory of Virtualbox itself (C:\Program Files\Oracle\Virtualbox by default) and select the iso file.




Before you jump into installing the Virtualbox Guest Additions, you should prepare your system as suggested by the Virtualbox documentation. To do this, run the following command:

# yum -y update && yum -y install gcc make kernel-devel
# reboot


Now from the command line in the CentOS guest mount the image as a virtual device:
# mount -t iso9660 /dev/cdrom /mnt

After this, run the set up:
# cd /mnt
# sh ./VBoxLinuxAdditions.run

Now you are supposed to install a package named dkms, but this is not available by default in the CentOS yum repo. Whenever you run updates on the kernel in your CentOS guest, you will need to run this command:

/etc/init.d/vboxadd setup


"Network is unreachable" CentOS 6.2 as guest in VirtualBox.

Since this is the first post here I will explain that I have started working on a project and am going to use this blog to document progress and post various problems I run into with the solutions that worked for me.

The project I am working on involves using multiple distributions of Linux, and since I don't want to multi-boot I am using VirtualBox to develop specific things for each distro. I am running Windows 7 Ultimate 64 bit on the host machine.

After I had installed CentOS 6.2 in a virtual machine I ran into trouble with connecting to the internet. I tried using NAT, bridged, and host-only options in the network settings for the VM and still no such luck. After some googling and banging my head on my keyboard trying to find a solution I had come up with nothing that had worked. I eventually came across a post in the Debian forums that gave a couple of commands that I punched into the shell, more specifically the route command.

The output of the route command was as follows:


Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.0.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
169.254.0.0 0.0.0.0 255.255.0.0 U 1003 0 0 eth0


... or something to that extent.
The issue was that I was behind a router and it wasn't reaching external networks. I was able to ping other VMs and my host machine and from both the guest and the host, but not able to ping google.ca. I would get a message telling me "network is unreachable".

The problem was solved when I changed the configuration a little bit in /etc/sysconfig/network-scripts/ifcfg-eth0. I appended the following lines to the file. I might also note here that when you configure your ifcfg file to work behind a router, the IP that you choose to use (I use a static internal IP for consistency, and it might be wise if you will be using them as virtual hosts) needs to be in the same network as your router. My router gateway is at 192.168.0.1 and this VM is using the static IP 192.168.0.19.

The entire file in it's working state is:


TYPE=Ethernet
DEVICE=eth0
BOOTPROTO=
NM_CONTROLLED=yes
NETMASK=255.255.255.0
USERCTL=yes
ONBOOT=yes
MACADDR=00:00:00:00:00:00
IPADDR=192.168.0.19
PEERDNS=yes
GATEWAY=192.168.0.1
NETWORK=192.168.0.0


After editing the fie and running /etc/init.d/network restart I was able to ping google and update/install things through yum.

That's what worked for me, if you've run into the same problem and were able to solve it in a different way please do let me know how you were able to solve it.

EDIT:
Upon setting up yet another VM and running into this same problem using Bridged Networking in Virtualbox, I was able to ping local machines, but not external machines on the internet or domain names.
To fix this issue and be able to resolve domains, I added this line to /etc/resolv.conf:

nameserver 8.8.8.8

After this I was able to resolve domains. There's probably another fix for that to do with the settings I configured previously, perhaps not. That's what worked for me and for those of you who don't know, 8.8.8.8 is one of the IPs for google's public DNS servers.