Nick Bolton's questions

I've just spent the better part of 4 hours trying to get Cacti set up on my Windows 2008 server (involves setting up PHP, MySQL, etc) and I've managed to get the UI working, but can't get the graphs working.

It's very clear to me that Cacti wasn't really designed for Windows (or at least, it wasn't designed to be easily configured on Windows). Surely there's something like Cacti designed specifically for Windows that uses MSSQL and ASP.Net? If this were true, then it would save me an incredible amount of time and hassle.

This is a weird one. Allow me to clarify, I have a Debian Linux file server - and I need to access some files from it. Usually, I would just use a network for this (wifi, ethernet, etc) but unfortunately the client computer can only access files over USB (the client is designed so that you can insert a USB flash drive, from which it reads).

Is there a way of making a USB drive pretend that it is a flash drive (or similar)? Perhaps that's phrased badly; to put it another way -- how do I make a USB port provide a file system instead of using it to access one?

Update: I forgot to mention that I'm using Debian Linux for the file system.

I see have an application which is using 40,000 handles. Could this be causing my system to become sluggish and unresponsive? What number of handles can Windows have before it becomes unstable?

My Windows PC is on two different networks, but I have programs that work better on a particular network. So for example, Outlook needs to use network A, but Chrome needs to use network B. I don't really want to have to add static routes every time I visit a new site, and there are some applications that still need to use the internet on network A.

What I think what I need is a program that can bind a program to a network. So that individual programs can use specific gateways... Is there such a program?

Even better, it would be great to have a local HTTP proxy which uses different gateways depending on the website I visit, but that might be out of the question...

I've tried to browse to .html, .png, .css, .txt files, but the page just comes up blank. I have tried browsing from other machines - same issue. Also (may be unrelated) there are no log files where I expect them to be: C:\inetpub\logs -- it only contains 1 empty wmsvc dir. Any ideas?

I'm using Network Solutions to host one of my domains for the sole reason that they [sometimes] allow me to change my glue records online. However, they keep changing their site and some times they remove the ability to edit glue records. Some times they just randomly turn off the feature and you need to ask them to turn it back on. I believe the link is called "Manage Host Servers". Quite annoying.

Is there a registrar that I can move to that will let me change my glue records without any messing around?

If I run the following command on Debian Lenny:

apt-get install git

... I'm unable to use git:

# git --version
bash: git: command not found

Possibly related to:

• SQL Server 2008 Install fails error reading etwcls.mof

After manually removing MSSQL 2008 from my system (uninstall failed to remove two instances), I receive the following error when trying to re-install:

The MOF compiler could not connect with the WMI server. This is either because of a semantic error such as an incompatibility with the existing WMI repository or an actual error such as the failure of the WMI server to start.

It seems that mofcomp is failing with one of the .mof files, but I'm not sure which, or why. Digging through the connect article gave some indications, but no solution. I've run winmgmt /salvagerepository, which returns "WMI repository is consistent".

Currently, I'm unable to install MSSQL 2008. Please help!

I'm trying to reproduce some issues on a deployed application where the MSSQL server and client are running in two separate machines. I think there may be network issues between the two machines, so I'd like to try and reproduce these conditions on two Hyper-V virtual machines (on the same virtual server). Of course, the network for these virtual machines is "local" so it's actually far from the conditions in a live environment.

Is there a program I can run on either virtual machine which will degrade the network performance? Or maybe any other work arounds? For example, one way to reproduce the conditions may be to run the VMs on separate Hyper-V servers in geographically dispersed locations (so the SQL traffic goes over VPN or something) -- but this is a little long winded I think. There must be a simpler way.

If a Windows server has several thousand ports open in LAST_ACK state (as shown by netstat), what could this mean? Is it because one endpoint is waiting for the other to respond?

There's a few similar questions which I tried to work out the answer from, but so far I have been unsuccessful. Please advise how I can always redirect http to https (and also remove www. from the hostname in the process). Also a side note, it would be nice to do this inside the main Apache conf rather than .htaccess - but I imagine this will not apply to most people.

• .htaccess redirect www to non-www with SSL/HTTPS
• Redirect subdomain request to subdirectory + https (using .htaccess)
• htaccess redirect for non-www both http and https

Update:

I've added this snippet to a VirtualHost section:

RewriteCond %{HTTP_HOST} ^www\.(.*)
RewriteRule ^.*$ https://%1/$1 [R=301,L]

... but it has no effect when I access http://www.domain (it should redirect to https://domain)

Update 2:

It had no effect because I did not use RewriteEngine on - so it works now:

RewriteEngine on
RewriteCond %{HTTP_HOST} ^www\.(.*)
RewriteRule ^.*$ https://%1/$1 [R=301,L]

I figured I'd share my question here and then answer, as there seems to be many people stuck in my position - but no definitive answer. The problem is, if you apt-get remove mysql-server, it does not clean up the configuration and database files, so if you've somehow screwed them up, then installing again, will not replace them. So there seems to be many people asking "how do I completely remove mysql-server, so that I can re-install a fresh?" -- everyone answers with apt-get remove --purge mysql-server -- I'm not sure why, but this does not fully uninstall. My answer follows...

After some digging and experimentation, I figured out how to install Python 2.6 on Debian 5.0 using apt-get. So I thought it best to post on SF as a question and answer. I know this can be manually built and installed, but many people prefer to use apt-get (including myself).

We have a directory of files (about 4GB) - and we'd like to synchronise between 3 locations across the internet. However, we're very conscious about bandwidth, so we need the following features:

• Schedule when synchronising should occur (e.g. 6pm - 8am)
• Specify a speed limit (e.g. 100KB/s)
• Don't synchronise when certain applications are running

We also need the following:

• Works on Windows and Linux
• Is free or not too expensive

Is there such an application in existence? We ruled out Dropbox because it's not free over 2GB, and Live Mesh because we can't set a speed limit or schedule.

According to a guide on the Linux directory structure, /usr/ is for application files, and /var/ is for files that change (I assume this means "files that belong to the applications"). Is this correct?

If this is the case then I'm a little torn between using either. A website is an application (if it's dynamic, so to speak), but in other cases it is just a collection of files used by Apache.

The default www dir lives in /var/www/, so should we follow suit by using /var/websites/ (or something similar), or choose /usr/websites/ since they could be applications?

This is a very trivial question, but it's bugging me nonetheless. For our case, I'm leaning toward /usr/web or something like that, since our websites are all applications.

Update:

This is for our company websites; it's not a shared hosting server, so we don't need to worry about separating them in /home/ or anything like that.

Each tutorial seems to have a different opinion on this. For my ISC BIND zones, should I use /etc/bind/zones/ or /var/cache/bind/? In the last install, I used /var/cache/bind/ but only because I was guided to do so; however I just spotted a pid file in there for this new Debian install, so I figured that using the "working directory" to store zone files probably wasn't the best idea. It seems that many admins use this so they don't have to type the full path when declaring a new zone.

For example:

file "/etc/bind/zones/db.foobar.com";

Instead of:

file "db.foobar.com";

Is obviously easier to type, but is it good or bad practice?

Some may also suggest setting the working directory to /etc/bind/zones:

options {
    // directory "/var/cache/bind";
    directory "/etc/bind/zones";
}

... but something tells me this isn't good practice, since the pid file would be created there I assume (unless it's just in /var/cache/bind by coincidence).

I took a look at the manpage but it didn't seem to say what the directory option was for, any ideas exactly what it was design for?

Here's some output that I hope is helpful:

nick@home-sv-1:~$ crontab -e
/var/spool/cron/crontabs/nick: Permission denied

nick@home-sv-1:~$ echo $EDITOR
emacs

nick@home-sv-1:~$ ls /var/spool/cron/crontabs
ls: cannot open directory /var/spool/cron/crontabs: Permission denied

nick@home-sv-1:~$ sudo ls -al /var/spool/cron/crontabs
total 12
drwx-wx--T 2 root crontab 4096 2009-10-25 20:45 .
drwxr-xr-x 3 root root    4096 2009-05-18 01:19 ..
-rw------- 1 root root     612 2009-10-25 01:20 root

Problem:

Crontab does not have setguid, and is not in the crontab group.

home-sv-1:~# ls -al /usr/bin/crontab
-rwxr-xr-x 1 root root 32048 2009-08-30 03:34 /usr/bin/crontab

Solution:

As root...

home-sv-1:~# chown root.crontab /usr/bin/crontab
home-sv-1:~# chmod g+s /usr/bin/crontab

home-sv-1:~# ls -al /usr/bin/crontab
-rwxr-sr-x 1 root crontab 32048 2009-08-30 03:34 /usr/bin/crontab

I always wonder this, every time someone gives me access to their server, I always want to find out what distro it is. I figured it'd be more fun to actually type a command rather than asking the admin.

Using my Django app, I'm able to read from the database just fine. When the application didn't have permission to access the file, it gave me this error:

attempt to write a readonly database

Which made sense. So I edited the permissions on the file, so that the Apache process had write permissions. However, instead of it being able to write, I get this cryptic error:

unable to open database file

If it's useful, here's the entire output:

Request Method: POST
Request URL:    http://home-sv-1/hellodjango1/polls/1/vote/
Exception Type: OperationalError
Exception Value:    
unable to open database file
Exception Location: /usr/lib/pymodules/python2.5/django/db/backends/sqlite3/base.py in execute, line 193
Python Executable:  /usr/bin/python
Python Version: 2.5.2
Python Path:    ['/var/www', '/usr/lib/python2.5', '/usr/lib/python2.5/plat-linux2', '/usr/lib/python2.5/lib-tk', '/usr/lib/python2.5/lib-dynload', '/usr/local/lib/python2.5/site-packages', '/usr/lib/python2.5/site-packages', '/usr/lib/pymodules/python2.5', '/usr/lib/pymodules/python2.5/gtk-2.0']
Server time:    Sun, 23 Aug 2009 07:06:08 -0500

Let me know if a stack trace is necessary.

I'm trying to deploy a Django application, but I've hit a brick wall. On Debian (latest), I've run these commands so far:

apt-get install apache2 apache2-doc apache2-mpm-prefork apache2-utils libexpat1 ssl-cert libapache2-mod-python python-django

I've tried adding the module manually in the Apache 2 config files, but to be honest I'm totally lost. It's totally different to Apache version 1 which I used years ago.

Syntax error on line 7 of /etc/apache2/sites-enabled/000-default:
Invalid command 'PythonHandler', perhaps misspelled or defined by a module not included in the server configuration

I've added the following to my sites-available/default file, between the tags.

    <Location "/">
              SetHandler python-program
              PythonHandler django.core.handlers.modpython
              SetEnv DJANGO_SETTINGS_MODULE hellodjango1.settings
              PythonDebug Off
    </Location>

Here's what tutorials I've used so far, without much luck:

• Django | How to use Django with Apache and mod_python | Django Documentation
• How To Install Django On Debian Etch (Apache2/mod_python)