Updated to Windows 10 killed my Debian GRUB (U)EFI Boot

After the Windows 8 -> Windows 10 Upgrade the Debian testing UEFI Grub2 only showed it's rescue shell. On Debian, I use full disc encryption and have a seperate boot partition - here is how I fixed my Grub without using some LiveCD or Live USB:

On the Grub rescue shell, enter the following - in case you don't know which partition holds what, you can also try > ls (hd0,gpt7)/ to get some content information. I knew that number 7 was my dedicated boot partition and number 8 the encrypted root file system.

XYZ? New domain name, small design changes - whats next?

If you're not the first time on my webpage, you'll notice that I've changed again the domain name of this place - now it's "nikolaus.xyz". A short FAQ about this change:

Why? The last years the domain name / URL of my internet presence was "nikolauspolak.info", that was surely ok (because it's my full name), but for me personally it was to long, too much characters to type. When last year the new domain endings showed up I thought that this is the right time to get a nice, suitable, short name - found this one.

OpenVZ/Proxmox Container: rsyslog problems after dist-upgrade

Right after a update of Ubuntu Server I faced that rsyslog used 100% CPU and that /var/log/syslog got filled up with following line, which was repeated almost every second:

 imklog: error reading kernel log - shutting down: Bad file descriptor

The problem is, that rsyslog tries to log kernel messages, which is as OpenVZ/Proxmox Container default not possible - to fix this, remove/comment out following line in the file /etc/rsyslog.conf (here already done):

Lighttpd 1.4.35 with PFS on Debian 7 wheezy

After I wrote the howto for Apache2 now a howto for lighttpd and PFS, where I couldn't use "apt-get build-dep lighttpd" because my installed MariaDB packages blocked some mysql-dev-lib packages. Tested on Debian 7 ECC keys and ECDH ciphers working. When you've already do some compiling and apt installing, it should be simple to follow:

1. Preperation - get source and needed tools

Apache 2.2 with PFS on Debian 7 wheezy and Ubuntu 12.04

During I was changing all ssl keys because of the heartbleed bug I thought it is a good time to enable PFS on my Debian 7 und Ubuntu 12.04 LTS servers (ECC keys and ECDH ciphers). Here is a step by step howto, when its not your first time to install something from source it shouldn't be that hard:

1. Preperation - download source and needed tools

Wrong display of free/used diskspace on OpenVZ/Proxmox containers

After I moved some OpenVZ-containers to a new server, the output of "df -h" and "du -sch /" was more different than I ever expected. I guess it was because of some old vzqota-files, because I used earlier existing VEID's.

The solution: Let the quotas get recalculated with following one-liner. Sorryly you've really to switch the OpenVZ-container for a short time offline:

# vzctl stop 199; vzquota drop 199; vzctl start 199

Replace the number "199" with the problematic VEID - now used&free diskspace should be displayed again correctly.

Using ownCloud to sync game savepoints and app data

I'm using this since some time - at most for OpenTTD, but it should work the same way for almost every game or app data you want to keep in sync on your desktops&laptops.

Tested as always only on Linux, 2 easy lines in the terminal and you're done:

[email protected]:~$ mv .openttd ownCloud/my-openttd
[email protected]:~$ ln -s ownCloud/my-openttd .openttd

Done on the first computer :-) On the second (and possible third, fourth, ...) please delete the directory if it exists and then repeat the second line of computer one:

apt-get update: procps ... start: Job failed to start

Seeing an error message during a server upgrade is something you don't want to see - but it seems many people are hit by that Ubuntu bug. Seems my systems were hit because I use OpenVZ (or, better said, Proxmox) for virtualizing my servers.

In the bugreport comments #5 and #11 you'll find the solution: comment out kernel.kptr_restrict = 1 or (re)move the file /etc/sysctl.d/10-kernel-hardening.conf

Now the update process can complete :)

Pages