lindelof's questions

I have an Ubuntu desktop (say orc) that runs a simple OpenVPN client. The OpenVPN server is called legolas.

I want all HTTP traffic coming from legolas to be forwarded to a machine, let's call it gimli, on the same local network as orc. I want orc to be the mediator between gimli and legolas. gimli is NOT running an OpenVPN server.

I've trid using FireStarter to configure orc as a pseudo-gateway, with the tun0 device acting as the local interface and eth0 as the internet device, but to no avail. legolas can ping gimli but cannot get any HTTP requests through.

Any idea how I should proceed?

Once upon a time I introduced the following line in our sources.list file:

deb http://ftp.ch.debian.org/debian/ testing main

I did that because I needed version 3.4 of the octave package, whereas the latest version in stable (lenny) was 3.2.

Fast-forward a few months, when I have completely forgotten about this change, and I decide to do a full upgrade of that server. I dutifully run

aptitude update
aptitude full-upgrade

And after a while, I discover to my horror that I have just upgraded to a testing distribution (wheezy).

None of the services running on that machine appear broken and everything seems normal. However, what should I reasonably do now? The machine runs only services used internally which, although not mission-critical, would be a pain in the butt if they were to fail.

The machine is hosted in a remote data center to which I have no access, and I am rather unwilling to go to the trouble of reinstalling a server from scratch if I can avoid it.

Note also that I am not a full-time sysadmin. Like any techie in a startup team, I am just the de-facto Linux guy.

What's the difference between useradd and adduser? When/why should I prefer using one or the other?

possible duplicate: Assessment & Planning Toolkit

I'm planning to rent a dedicated server from our provider that will host several services (SVN, LDAP, etc) each as a dedicated virtual machine. I guesstimate there will be between 5-10 servers, none of which should be under a particularly heavy load.

If I run this under Xen, how much memory/processing/disk should I ask for? Roughly? Any rules of thumb?

We're a small startup and I need to setup several applications, almost all of which require a database.

I face the choice of setting up one instance of MySQL, with one database for each of our applications, or of setting up a separate instance of MySQL for each application.

What is the generally accepted best practice in this situation? One database to serve them all, or divide and conquer?

I'm completely new to Cfengine and I've just installed it on my Debian system. But when I try to invoke any of its executables as a non-privileged user, here is what I always get:

$ /usr/sbin/cf-key 
Fatal cfengine error:  !!! No access to workspace

Strangely enough, it looks like my local workspace got correctly created:

$ tree .cfagent/
.cfagent/
|-- bin
|-- lastseen
|-- outputs
`-- reports

So any idea what I'm doing wrong here? Googling for this error message drew a blank...

Everybody seems to agree that granting SSH login to the root user is a Bad Idea (tm). But does this mean I should create an account for myself on every machine that I administer? One with sudo privileges? Do you usually create one generic account, or do you create one account for every mammal that's supposed to access that machine?

I've often seen system administrators download and install server applications, including "stock" applications such as Apache, under a dedicated /srv or /opt directory instead of installing them from the official repository, e.g. apt-get install xxx.

Is it really a bad idea to install and run a webserver, or an email server, or even less common services such as Zope from the distribution's official packages?

I have WebLogic 11g installed on my machine under /srv/wls, and I start Node manager on it with /srv/wls/wlserver_10.3/server/bin/startNodeManager.sh.

I create my domain with config.sh, and then I want to start the Admin Server of that domain with Node Manager. So I open a WLST session and do the following:

> nmConnect(...)
> nmStart('AdminServer')

But when I then visit localhost:7001/console, and login, the screen on the browser remains blank and the server logs tell me the server has run out of PermGen space.

I don't encounter this problem when starting the Admin Server through the shell script under $DOMAIN/startWebLogic.sh. Then it works fine.

Any idea what I am doing wrong? I couldn't find anything in Google nor on the Oracle forums...