Any reason to Upgrade from LMDE install to new '(201012)'?

[chris_debian]

Hi, all!

I'm really pleased that the LM team have managed to get the '(201012)' release 'out of the door' for Christmas. I'm downloading the 32 and 64-bit torrents so that I can seed and try to help out with what I expect will be a massive demand for the ISOs. Nice work LMDE people!

When the initial LMDE release was published, I eagerly downloaded it and have an (xfce) installation on my main machine, which I am very happy with.

0. Is there any reason that I should install using the new image, given that I am happy with my existing install and all the updates have been applied?

I note from the latest 'release notes', that the famous '“4 lines of code better than 200″' kernel patch will be available in this new release, but see in my existing installation that a kernel upgrade is available (if I run apt-get upgrade), but the 'MinutUpdater' is holding it back. I could run apt-get upgrade, but I know from previous trial and error, that this will upgrade my XOrg and break it (ATI/ Radeon issues that are not LMDE specific).

1. Is it possible to get the kernel update, without upgrading X (ie, running apt-get upgrade)?

I'm blissfully happy with my installation, so if the answer to question 1. is 'no' then this is not really an issue, just wondered . Any thoughts on question 0.?

Cheers, all!

Chris.

[drzoo2]

I was just going to ask the same question. After reading some info on this I'm confused as to when the kernel patch was rolled out.

z

[kwevej]

no reason

..for the kernel, I'd recommend Liqorix kernel which has a better scheduler
if u prefer the standard kernel:
- that 4 lines of code is not an update that comes with kernel, it is just an modification of .bashrc file

[drzoo2]

Ok, A google search found this.....

Code added to .bashrc

Code: Select all

if [ "$PS1" ] ; then 
   mkdir -m 0700 /sys/fs/cgroup/cpu/user/$$
   echo $$ > /sys/fs/cgroup/cpu/user/$$/tasks
fi


and then execute this mount and directory command.....

Code: Select all

mount -t cgroup cgroup /sys/fs/cgroup/cpu -o cpu
mkdir -m 0777 /sys/fs/cgroup/cpu/user


I assume the mount has to be executed at login via a script?

[chris_debian]

Thanks for the replies, guys! I thought the kernel thing was an official kernel.org release. My LMDE box is sufficiently fast that I do not need to mess with .bashrc, etc.

I'll stick with the LMDE build that I already have.

Thanks, again.

Chris.

[tdockery97]

I could run apt-get upgrade, but I know from previous trial and error, that this will upgrade my XOrg and break it (ATI/ Radeon issues that are not LMDE specific).

Cheers, all!

Chris.

I also run ATI/Radeon and had my system borked from updates. Several Debian veterans had previously said that you should not do upgrades with X running. Of course I never listened, and of course my system was repeatedly borked. Initially I couldn't drop out of X to do updates because of course Network Manager immediately shuts down. Recently I installed WICD which does come up even outside of X. So now to do updates I just ctrl-alt-f1 to exit X to command line and do dist-upgrades from there. There have been several xorg updates and my video has not borked once. Just something you might want to try.

Tom

[libssd]

I have what I hope is a related question. After running LMDE on an 8gb SDHC card for several months, I decided to reinstall from the image that was available around Dec 15. No known problems, and I have applied several updates since then. However, when I tried to install updates on Dec 24, I get the following error message from the update manager:

Could not apply changes!
Fix broken packages first.

However, Synaptic reports zero broken packages.

Since this installation is my sandbox, a fresh install from the latest LMDE distribution is an option.

Suggestions?

Edit: I found an answer in another thread:

Code: Select all

sudo apt-get update && sudo apt-get upgrade && sudo apt-get dist-upgrade

But the question remains: why was I getting this error message?

[drzoo2]

Thanks for the replies, guys! I thought the kernel thing was an official kernel.org release.

Me too but I have not heard or read anything official on this. I would really like an admin to post a thread on this subject.

My LMDE box is sufficiently fast that I do not need to mess with .bashrc, etc.

.bashrc is a great way to use alias's. I have a slew.

ex: when I type sshserv it aliases to my server address. Sure beats ssh -X 192.168.0.240

z

[kwevej]

But the question remains: why was I getting this error message?

because:
http://blog.linuxmint.com/?p=1595
viewtopic.php?f=141&t=62201&p=357525&hilit=broken#p357525

btw: I was thinking about the same - running LMDE from an SD card. Which filesystem do u use for that?

[libssd]

But the question remains: why was I getting this error message?

because:
http://blog.linuxmint.com/?p=1595
viewtopic.php?f=141&t=62201&p=357525&hilit=broken#p357525

btw: I was thinking about the same - running LMDE from an SD card. Which filesystem do u use for that?

Re the error message, I had already manually updated to the new Mint desktop.

Re filesystem, ext2 or ext4 are supposed to be the fastest, BUT if ext4, several tweaks are recommended to minimize disk writes. Here are a couple of references:

http://samiux.wordpress.com/2009/06/09/ ... -disk-ssd/
http://blog.bodhizazen.net/linux/netbook-optimization/

Here's what I did:

1. sudo gedit /etc/fstab Add noatime parameter to the line that defines your root partition:
UUID=31b8c0a8-db49-4835-b7f0-2c1f4392ea92 / ext4 noatime,errors=remount-ro 0 1

2.Mount /tmp /var/log and /var/tmp in RAM. Add these lines to the end of fstab to mount /tmp (temporary files) as tmpfs (temporary file system):

tmpfs /tmp tmpfs noexec,defaults,noatime 0 0
tmpfs /var/tmp tmpfs noexec,defaults,noatime 0 0

3. sudo mount -oremount /dev/sda1

4. Reboot

Other tweaks are suggested in the articles I referenced. With LMDE, I found no difference with any of the three I/O Scheduler options. YMMV