This site is now 100% read-only, and retired.

XML Logo

Posted by eric on Tue 23 Nov 2010 at 08:47

Hi all, I'm looking for a syslog server to centralise our many (mainly Windows) servers. There seems to exists two real alternatives: rsyslog and syslog-ng. A comparison (from 2008) exists here: http://www.rsyslog.com/doc/rsyslog_ng_comparison.html but I prefer to ask here for real experience.
So my questions are:

  • what syslog server are you using?
  • have you compared syslog-ng and rsyslog features before choosing one? and when?
  • do you feel there are limitations, maybe blocking limitations in the one you use?
  • what do you use (if you use any) web interface for visualization? Do you any aggregation/correlation software for your logs? And which?

As you see, my idea is to use a syslog server to centralise logs, but also to provide a web interface for visualisation and even more if i found great software.

I'm waiting for your comments :)

 

Posted by eric on Fri 23 Oct 2009 at 15:53
Tags: none.
Hi, I finally decided to install and manage my own mail server. All is good with a combination of Posftix, Dovecot, sieve filters, etc.
For the moment, I'm testing on a test domain but before I switch all my real domains, I have a problem: what if my server crashes and i'm not online (sometimes i'm on vacations away from internet)?!!
I need a backup mx...

But for the moment my searches returns nothing about backup mx, free (or affordable). Do you know any service?

I know running a backup mx is not a simple task, and it's better not have one than a bad one, but... I think I can allways filter all the mails coming from the backup mx on my managed one, and my main problem is disponibility.

 

Posted by eric on Wed 16 Jul 2008 at 19:07
Tags: none.
I'm currently evaluating Redmine as a project manager and, to keep my virtual server 'clean', I have chosen the 'rails+apache+mod_fcgi without mongrel' solution.
But the application is very unstable: it returns errors (apache logs says something like 'premature end of script: dispatch.fcgi') frecuently and only returning to the base url (http://test.domain.tld) I can return to the application.

So my question is: do you know if it's normal (i don't think so) and do you think installing it using Mongrel (like it's indicated in the vast majority of tutorials) is the solution of this problem.

I ask this because howtos about rails app installation without mongrel are hard to find, and nearly all talks about Mongrel (a Ruby webserver) in combination with apache mod_proxy.

 

Posted by eric on Mon 14 Apr 2008 at 14:59
Tags: ,
On a Etch installation, I'm trying to generate certificates and a root certificate (to integrate in a Active Directory environment). I'm using the guide available here (http://www.debian-administration.org/articles/284), but have no success: the first real command (creation of root certificate) never resume.
Anyone have an idea? Maybe, this guide is outdated?

 

Posted by eric on Wed 23 Jan 2008 at 16:04
Tags: none.

Hi all,

I'm currently configuring a Xen server with several domUs, but all of them sharing a single IP address. I use xen-nat and some iptables to redirect ssh and http ports [1].

So i'm currently connecting with ssh to each domU with commands like
$ ssh root@IP -p 1022
using 1022, 1023,... redirected to each domU port 22.

The problem is ssh key verification that force me to erase my .known_hosts IP key each time i connect to a different domU.

Do you know if there is a way to disable completely this behaviour or (better) only for my IP host?
(i have alreadly checked StrictHostKeyChecking but it don't seems to work because each time it cancel the connection)

[1] check http://blog.sietch-tabr.com/index.php/post/2007/07/10/Xen-NAT for more details (in French)

 

Posted by eric on Tue 16 Oct 2007 at 18:14
Tags: , , ,

As my CPU is now HVM compatible, I tried to install Windows on Xen, and got it working after some days searching and searching on the internet the good combination for my conf file.
If someone is interested I can post details here.

:eric:
http://blog.sietch-tabr.com

 

Posted by eric on Tue 11 Sep 2007 at 21:27
Tags: none.

nspluginwrapper is now in Debian testing, and installing it enable Flash plugin for both Mozilla Seamonkey/Iceape and Firefox/Iceweaseal (using Ubuntu doc, I got only Iceape working) so it is great!

:eric:
http://blog.sietch-tabr.com

 

Posted by eric on Fri 31 Aug 2007 at 19:18
Tags: none.

I finally rechecked my Xen+nVidia binary drivers problem, and this time solved it! Using the nvidia-kernel-source and nvidia-glx packages from unstable: the 100.14.11-1 version, and using the patch available here, I achieved to compile the module and got it running... ;-)

:eric:
http://blog.sietch-tabr.com

 

Posted by eric on Sun 29 Jul 2007 at 21:17

Ok, for the moment, I let aside the nvidia binary problems using the free nv driver, and try to make flash and other 32 bits apps running on my pure amd64 system.
I have checked all the d-a.org articles about the problem and a few others on the internet.

I think the chroot solution is a little bit complicated (management of two systems, problems sharing files, disk usage) so I looked to the solution indicated by ajt here. I ever found the getlibs script (see here) to simplify my dependencies problem, but I can't understand how i can run the flash player for Firefox as a plugin.
If i check with ldd the libflashplayer.so file, i see that no dependency is missing. But... the files needs to be put in the firefox plugins user directory (or in /usr/lib/firefox/plugins for systemwide), the application is not standalone and my Firefox is 64 bits, so when I browse a flash-enabled page, it simply crashes !

I there a solution for that? Or do I need to run a 32bit Firefox and so not use Debian package management?
I'm a little bit lost with these library problems, I don't really know/understand the problems and don't want to make mistakes, and want to have a the simplest possible solution, can anyone help me?

:eric:
http://blog.sietch-tabr.com

 

Posted by eric on Wed 25 Jul 2007 at 19:29
Tags: , , ,

So, i successfully passed the Nvidia driver phase. No real problems with the standard amd64 kernel (see below for more...), there's a lot of documentation on the web to use a nvidia graphic card with the proprietaries drivers (don't forget the nvidia-glx package!). The only thing is the dual-head functionnality that i can't really try without a second screen (the KVM output an 'out of frequency' error on the screen on one of the VGA/DVI outputs).

Now, comes the Xen part. It drives me crazy, i rebooted multiple times but i finally found the right combination in my /boot/grub/menu.lst to boot the linux-image-2.6.18-4-xen-amd64 kernel on my ASUStek M2N motherboard. Here it goes: (relevant parts in bold)

title Xen 3.0.3-1-amd64 / Debian GNU/Linux, kernel 2.6.18-4-xen-amd64
root (hd0,0)
kernel /xen-3.0.3-1-amd64.gz noapic
module /vmlinuz-2.6.18-4-xen-amd64 root=/dev/mapper/vg--root-lv--root ro acpi=off noapic nolapic console=tty0
module /initrd.img-2.6.18-4-xen-amd64
savedefault

And finally, the problem comes on the Xen+Nvidia phase. I compiled my kernel modules for Nvidia the same way i did it with the standard kernel, but everytime i launch X (via startx or gdm) the screen turn blank and the pc reboot... with no messages, neither on the screen, neither on my logfiles... i'm a little bit lost, so if anybody got an idea...

:eric:
http://blog.sietch-tabr.com