Category: Linux

Our friends at openwrt released the Whiterussian RC5
http://forum.openwrt.org/viewtopic.php?id=4977

A lot of bugfixes….. Time to do a backup of my WRT54GS and upgrade

Time to upgrade to FC5…

First we upgrade our system to the latest FC4 updates:
yum upgrade

Upgrade the fedora release:
rpm -Uvh http://download.fedora.redhat.com/pub/fedora/linux/core/5/i386/os/Fedora/RPMS/fedora-release-5-5.noarch.rpm

Upgrade the yum:
yum upgrade yum

And now time for the full system upgrade:
yum upgrade

Wait……….do a reboot and you are up and running.

Fedora core 5 released today.

bofh waits for it upgrade, i will give it a couple of days for the mirrors to sync and then i will upgrade through yum

One common problem with the default settings of bash or the defaults settings of your bash_profile ~/.bashrc is the bash_history.

When you logout from your shell all the commands you entered there are writteen to the bash_history. Hmmm what happens when you logout from 2 or more shells ? Yes the last history list is written = you loose what you typed in the others shells you closed.

Instead of overwriting the history list to the history file you can append it.

shopt -s histappend
export HISTSIZE=4096

We also change the default HISTSIZE from 500 lines to 4096

Add the commands you your ~/.bashrc ~/.bash_profile and you are up and running

My main problem all the time is that i have small capacities on hard disks (8GB,10GB) and so on.

I don’t want to throw them away but i don’t want also to leave them lying around.

This time i did an exception: i bought a big one, a Western Digital WDC WD2000JB (200GB) plus this time i didn’t the same mistake => That is create partitions with knowning how big it should be and how you will resize them on the feature.

So i desided to give a try to LVM. LVM isn’t as difficult as it sounds. Follow the howto and you will be up and running in seconds.
Initialize the disk

[root@bofh ~]# pvcreate /dev/hdc
Physical volume “/dev/hdc” successfully created

Create a volume group
[root@bofh ~]# vgcreate volume_group /dev/hdc
Volume group “volume_group” successfully created

[root@bofh ~]# pvdisplay /dev/hdc
— Physical volume —
PV Name /dev/hdc
VG Name volume_group
PV Size 186.31 GB / not usable 0
Allocatable yes
PE Size (KByte) 4096
Total PE 47695
Free PE 47695
Allocated PE 0
PV UUID xWttvY-ruQD-AaiC-ibkp-SPk1-v5nZ-GMLXVY

Creating a logical volume
[root@bofh /]# lvcreate –size 10GB volume_group -n backups
Logical volume “backups” created

Create a filesystem on the /dev/mapper/volume_group-backups mount it and you have finished

When i will need more space for my backups i will do a lvextend extend the filesystem and that’s it.

Plus i can attach other harddisks on the lvm anytime i want 🙂

I am a fun of reiserfs. Period.

The only filesystem i had without reiserfs was on my laptop. 20GB of data on ext3.

I wanted to converted to reiserfs without doing a tar create a new filesystem on the disk and untar the data. Since i had a working backup set, i give a try on convertfs (http://tzukanov.narod.ru/convertfs/)

Boot from a CD with builtin kernel support for the filesystems you want to convert from and to.

Build the convertfs utilities on another machine and start the conversion.

Warning: It takes a lot of time. On my laptop (Pentium M 1.3 GHz) took about 2 hours to do the conversion.

Change your fstab and boot in a your new filesystem.

Everythink worked as expected.

Case closed.

Debian team released the new installer for the etch.

It appers that the next time (in about 1-2 years 🙂 that you will install Debian you will have a graphical version of the installer.

For the rest of us, an apt-get dist-upgrade will be sufficient 🙂

Also they dropped support for installations using 2.6 kernel for the sparc32 architecture…..

Time to buy a sparc64 ?

I don’t know what is going on on Digium but i must admit i cannot follow their releases anymore

Have a look on the announcements

Asterisk 1.2.0-beta1 29 Aug 2005
Asterisk 1.2.0-beta2 1 Nov 2005
Asterisk 1.2.0-rc1 9 Nov 2005
Asterisk 1.2.0-rc2 12 Nov 2005

It appears that we will have a stable 1.2.0 till the end of the month.

I will do some tests in my production voip with the 1.2.0-rc2. The machine still runs on 1.2.0-beta1 and i want to upgrade it but i have no spare time…

First mail in my inbox today…. The release of Asterisk 1.2.0-beta2

A ton of bugfixies. Will i find some time to upgrade my asterisk installation ?

As a start i put it as the first item in my TODO list

Hmmm, a kernel release…
So ? Just another kernel release ?
Should i upgrade ?

No, wait for a few days. Everything works ok with 2.6.13

These are the usual questions after a kernel release 🙂

This time since it was a weekend i want to give a try.

The ipw2100 make it way to the kernel tree. Nice no more another compile for my wireless laptop network card to work.

The vesafb-tng patches works ok with the new kernel. Nice i will not miss my 1024×768 terminal 🙂

The suspend2 team hasn’t release a new version so it will stay a couple of days ( maybe a week) with out hibernation.

The vmware needs the usual vmware-any-any-update 🙂

It took me a couple of hours but everything is running smoothly.

Nice work kernel guys

I came across with a really nice utility for debian.

It is the deborphan. What it does ? It looks for orphaned libraries in your debian system. That is library packages that you no longer use.

Give it a try
[url]http://packages.debian.org/stable/admin/deborphan[/url]

Dag hasn’t provided yet an rpm of the openvpn 2.0 for Fecoda Core 4.

Here is one, just builded a couple of minutes ago 🙂
[url][/url]

I am using the Talisman firmware on my WRT54GS for about 2 months.

The last 2 days i was having problems with the connection to the internet. I am not sure that it was a problem with my adsl provider but i could connect at all 🙂

On the dmesg i was seeing messages like the following
[b]do_page_fault() #2: sending SIGSEGV to hotplug for illegal read access from
00000000 (epc == 00000000, ra == 00000000)[/b]

Nice ?

Rebooting and/or flashing again didn’t solve it.

I decided to see if it was a firmware problem and i switched to openwrt.
Openwrt released the whiterussian on 25-Jun-2005 and i can say that i am really enjoy it now, a lot of packages and a small, clean firmware

What i hate after upgrading a kernel ? Compile all your modules that there are not part of the main stream kernel.

If you use a custom kernel is easy. But if you want to stick for a reason with a prepackaged kernel sometimes is difficult.

The debian way = Allways build a package for your custom modules.

Download and install module assistant (http://packages.debian.org/stable/devel/module-assistant). Module assistant makes the installation of modules such the nvidia, ipw2100 peace of cake.

It will download the sources using apt, compile them and finally create a deb package.

So simple 🙂

After upgrading to kernel 2.6.12 with software suspend2 the ipw2100-1.1.0 driver stopped compiling.

The following patch take care of it. It solves the[b] is_multicast_ether_addr[/b] and [b]create_workqueue[/b] errors.

diff -Naur ipw2100-1.1.0/ieee80211.h ipw2100-1.1.0.test/ieee80211.h
— ipw2100-1.1.0/ieee80211.h 2005-03-14 18:54:24.000000000 +0200
+++ ipw2100-1.1.0.test/ieee80211.h 2005-06-21 11:48:06.000000000 +0300
@@ -643,10 +643,12 @@
#define MAC_ARG(x) ((u8*)(x))[0],((u8*)(x))[1],((u8*)(x))[2],((u8*)(x))[3],((u8*)(x))[4],((u8*)(x))[5]

+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,12) extern inline int is_multicast_ether_addr(const u8 *addr) { return ((addr[0] != 0xff) && (0x01 & addr[0])); } +#endif extern inline int is_broadcast_ether_addr(const u8 *addr) { diff -Naur ipw2100-1.1.0/ipw2100.c ipw2100-1.1.0.test/ipw2100.c --- ipw2100-1.1.0/ipw2100.c 2005-03-14 18:54:25.000000000 +0200 +++ ipw2100-1.1.0.test/ipw2100.c 2005-06-21 11:48:06.000000000 +0300 @@ -6401,7 +6401,7 @@ INIT_STAT(&priv->fw_pend_stat);

-#ifdef CONFIG_SOFTWARE_SUSPEND2
+#ifdef CONFIG_SUSPEND2
priv->workqueue = create_workqueue(DRV_NAME, 0);
#else
priv->workqueue = create_workqueue(DRV_NAME);

The machine that holds my blog, holds also everything for my domain 🙂

It was running fedora core 3 and it hasn’t been switched to Debian sarge (yet ?) due to the reason above.

After the release of Fedora Core 4 i had 2 options: Download the CDs and use them for the upgrade or do it over the net as i have done it for all my debian machines.

The isos are about 2.5 GB….:(

Upgrading over the internet using yum is simple.

Just follow this howto
[url]http://www.brandonhutchinson.com/Upgrading_Red_Hat_Linux_with_yum.html[/url]

The upgrade run with out any serious problems. The only problem after the upgrade was the php-mmache which i will investigate it a bit cause it gave me signal 11 error on the apache processes i suspect that was due to the fact i haven’t clean the php-mmcache cache directory.

A couple of reboots to test the new release and we are up and running

Finally after 3 years the new debian sarge waits for us in the mirrors.

For as that we were using testing upgrading is simple:
Change the sources.list to point to sarge instead of testing and do
aptitude update && aptitude -f –with-recommends dist-upgrade.

I had allready done it when the sarge was frozen last week :). No packets changed 🙂 🙂

In less than 4 months, 5000 definition where added into clam antivirus.

35000 as of today and counting.

The Register has an article about the decision of Debian release manager Steve Langasek to drop some architectures… Sparc is one of them.

[url]http://www.theregister.co.uk/2005/03/14/debian_reduced/[/url]

Could somebody please explain me what i i will run on my sparc machines (beside Solaris of course) ?

Let’s hope that they will reconsider it again….

Or maybe it is 2.6.11.1, or it is 2.6.11.2 or it is 2.6.11.3 ?

Edit: 14/05/2004 I think today they desided that the latest 2.6 kernel is the 2.6.11.3 🙂

I was thinking to upgrade my machines…. NO, I will stay with the 2.6.10.
Read why…
[url]http://kerneltrap.org/node/4827[/url]

Remember the kernel numbering has changed….
The *.x.y kernels are unstable.
The *.x only kernels are stable.

That means that 2.6.12 will be a stable one, and 2.6.11.y is unstable

Good or bad… Time will tell