As an example we are going to resize /dev/sda1 from 200G to 50G. Since the partition is the primary /root we need to use a rescue disk to boot the system, I used BackTrack from a USB stick (Howto:USB Stick).

1. use tune2fs to remove the journal from your ext3 partition:

   tune2fs -O ^has_journal /dev/sda1

   Now the partition is effectively an ext2 file system.

2. use “resize2fs /dev/sda1 50G” to resize the file system.
3. use fdisk to resize the partition: delete the old partition (no data will be lost! ). Create a new one of the desired size (exercise caution see below). Save changes.
4. use “resize2fs /dev/sda1” (no size this time) to resize the file system to the maximum available.
5. use tune2fs to add the journal agai:

   tune2fs -j /dev/sda1

   This turns the partition back to ext3.

Regarding the new size for the partition, it is important to allocate enough physical space to support the file system. I used the formula recommended by [2]:

We multiply the amount of blocks from the resize2fs output (1536000) by the size of a block (4k), and to go sure the partition is big enough, we add 3 to 5% to it (3% was enough for me, but if you want to go sure take 5%):

1536000 * 4k * 1.03 = 6328320k

The interesting number is the first one, and you can get it by looking at the output of resize2fs on step 2. You just need to specify that number when asked by fdisk (step 3) for the last cylinder of the new partition. Again from the same reference:

Last cylinder or +size or +sizeM or +sizeK (1-1247, default 1247): +6328320K

Note that this is not the vaule I used (I forgot to write it down), but I guess that this number depends on the hardware and the important bit is to learn how to apply The Formula.

References

This post consists of 100% recycled information, credit goes to:

• [1] Another resize ext3 problem
• [2] How To Resize ext3 Partitions Without Losing Data

Last minute note

Support for ext3 was added to resize2fs in version 1.19, more than 7 years ago. There is no reason to convert to ext2 first unless you are running a REALLY old system.

So you may avoid steps 1 and 5 if your resize2fs supports ext3.

What I want to do is to be able to shut down my computers when de charge of the UPS is below a certain level. So I have spent some time this afternoon configuring nut (network UPS tools).

From their website:

The primary goal of the Network UPS Tools (NUT) project is to provide reliable monitoring of UPS hardware and ensure safe shutdowns of the systems which are connected.

So it’s perfect for our purposes. It hasn’t be the first time that I have tried to configure nut, but It has been the one that I have finally managed to do it. It was as easy as finding mge and nut how-to.

In this first article I am going to explain how to configure nut to detect your pulsar ellipse ups. maybe someday I will explain how to set up the safe shutdown proccess.

The steps to make nut work are as follows:

1. identify the NUT driver needed: go to supported upss and find what driver you should use for your ups. newhidups does the job for my ellipse 1200
2. install NUT packages: apt-get install nut-usb usbutils. Once dependencies are resolved you end up installing also nut and hotplug packages.
3. configure NUT: see below.
4. test your configuration: issue upsc ellipse@localhost if no error message is returned, you are done.

You have to do a few tricks in order to set up the configuration files of nut. Create a copy of the configuration files from the examples dir: cp /usr/share/doc/nut/examples/* /etc/nut.

Then modify /etc/default/nut to look like this:

# start upsd
START_UPSD=yes

# start upsmon
START_UPSMON=yes

You have to configure /etc/nut/ups.conf, adding something like this:

[ellipse]
        driver = newhidups
        port = auto

Nothing to be done with /etc/nut/upsd.conf.

Add the following lines to /etc/nut/upsd.users file:

# Supervision user
[admin]
        password = mypass
        allowfrom = localhost
        actions = SET
        instcmds = ALL

# Protection user
        [monuser]
        password  = mypass
        allowfrom = localhost
        upsmon master

Reload the diferent services: invoke-rc.d hotplug restart and invoke-rc.d nut restart

Please take care to change the passwords given here because of obvious security reasons.

And as usual, don’t forget to google on it, browse through man, and read the official how-to.

You need some tricks to make it work.

You have to create an empty partition of 700 Mb (using repartitioning tools such as gqparted or the no-gpl Partition Magic).

Boot your linux, then format the new partition (/dev/hda6 in my case) using mkfs.ext2 program:
mkfs.ext2 /dev/hda6

Here is how my drive looks:

Mount it where you want, for example in /mnt/. Once you have it, let’s copy the contents of the Auditor’s iso image to that partition.

First of all get an iso image from Auditor’s website. Once you have it you need to mount it in your current linux filesystem, for example in /cdrom/ usign (as root) the following line:

mount -o loop /tmp/auditor-200605-02-no-ipw2100.iso /cdrom

NOTE: for mounting .iso files you have to had some features compiled in your kernel. These include:

#kernel options for .iso file mounting
Block Devices->Loopback device support
Filesystems->CD-ROM/DVD Filesystems-> ISO 9660 CDROM file system support

To copy the contents of the .iso file to the newly created partition use:

cp -a /cdrom/* /mnt/

To make it clear: now you have a partition in your hard drive that contains just the same files and folders of the Auditor’s cd-rom. Let’s boot from this partition. As boot manager I use grub so here is my customization of the menu.lst file:

# For booting the Auditor distro from /dev/hda6
title  GNU/Linux - auditor
kernel (hd0,5)/isolinux/vmlinuz root=/dev/hda6 ramdisk_size=100000 init=/etc/init lang=es splash psmouse.proto=imps dma nodhcp noapm noagp screen=1024x768 vga=791 nowheelmouse nomce  quiet BOOT_IMAGE=/KNOPPIX/KNOPPIX fromhd=/dev/hda6
initrd=(hd0,5)/isolinux/miniroot.gz