At work we recently acquired an Areca raid controller. Unfortunately the ftp site of areca is terribly unreachable for some days now. So i'm trying to mirror the site at our ftp site. If anyone has some stuff to contribute on Areca sata raid controllers please inform me! So I can add that to the mirror.

Update It seems Areca has switched to another server for their FTP site which seems to work. As soon as it started to work I ofcourse ripped there entire site and mirrored it at the above URL. Happy leeching!

By the way our Areca performs quite nicely in our RAID5 setup with 4x a Maxtor 300GB SATA disk. It has a read performance of 155 megabyte/sec and a write performance of 134 megabyte/sec (as measured by bonnie++). Quite nice :)

This is my public response to Nicole Braats paper on Cookies 'Cookies: Sweeties, Smarties of Monsters?'

My first response to the paper is that data stored in cookies could also be stored on an alternative way like in a database server at the webservers end. However cookies can reduce the size of possible users. If an database server stores data of a certain user based on an IP address the userbase can be the following types 'single user, single pc, proxy server or nat ip'. The cookie singles this domain to 1 element which is 'profile of webbrowser'. The profile of an webuser usually means 'single user or single pc with only 1 user'.

The cookie mechanisms intended use is storing temporary volatile data that usually only will exist one session with the webserver.The paper sets as a conclusion that cookies are infringement of privacy, however this is not true, if this was true then every webserver hosting any content would be an infringement of privacy as these all log at a minimum the following attributes of a webhit: ip address, requested file, useragent, time of request, size of request, (error)code of request.

Usually this information generates more info then a cookie can and this information is available as soon as a user visits a webpage. A cookie has to be explicitly set by the webserver on the clients side. Furthermore the test set that has been used to write this paper was only 53000 cookies collected from a small set of users that all fall in the same class. It's not a diverse research paper. In my opinion the paper is not something to expect from a University graduation paper.

Why I made this response to this paper is that all kinds of media are urging people to turn off cookies. This will generate a quite large impact for users as a lot of sites won't work anymore because the cookies are used to store temporary data which the website uses to identify yourself (for example a webmail server). Next to temporary data cookies are also widely used to store preferences of users on their local machine instead of on the server. This dumps quite a lot of overhead on a website, and also reducing overhead for users because they have to login on a site and remember passwords and logins for little non-private information like nr of items on a page and language.

This response is still a draft, The draft has been written at a moment I was not completely actively thinking about it, it was the first that popped up in my mind.

When I used Movable Type as my weblog tool I started to get very frustrated due to Comment Spam. I moved to Pivot for other reasons then the comment spam, but the nice side effect of pivot was that I also had no more comment spam on my weblog (not exactly because of pivot, more of the fact that comment spammers seem to like MT).

However with Pivot, referer spam seemed to arrive. Main reason looks to be that we have this referer list on the frontpage. And ofcourse spammers love publicity. Pivot handles this not really nice in my opinion. There are options within Pivot to block certain IP adresses and urls, but if you block these when you are already hit by them they will still show up in your last referer list on your front page. So basically harm is already done. Someone has a solution or a patch for this ? (I'm to lazy to read documentation at the moment concerning this)

UPDATE: Wow just when I saved this I saw a news item on my pivot admin site that Pivot 1.22 has somethings to prevent comment and referer spam. Let's upgrade!

Apparently a Telemarketeer of BBeyond called up the office today. Seems he wanted to sell a business SDSL connection. Unfortunately they called the number of the frontdesk of our mother company and not our number *grin*. What would we need a SDSL connection for if we have this shiny, fast and stable E3 connection from Essent. that goes straight to our own network in Amsterdam.

I hope someday another ISP calls and tries to sell us an Internet connection. I love to talk to them. Other Telemarketeers get the ZapaTeller() or the telemarketeer torture (somebody in for translating it to dutch ?).

UPDATE: Guess what, we received snailmail spam from Tiscali today. trying to sell us Office SDSL! Fortunately for them all their telemarketeers where busy, so we couldn't buy it.

Sjeemz his Peering-Slut is a wonderful PHP script designed to check which peerings you currently do not have on the AMS-IX. Unfortunately it will only work for Juniper routers. So I decided to write a little perl script that converts cisco/foundry configuration files to XML that can be parsed by peering-slut.

Ofcourse I delivered it to the public domain, and it can be found here.

As we are now almost a week further one using Asterisk as a PBX at work here are some more hints,tips and patches that can help everyone with asterisk:

- Asterisk and ISDN4Linux hate each other concerning CallerID. Asterisk excepts an empty or non-existant callerid as a call without CallerID. ISDN4Linux will always send 0 to indicate an absent CallerID. Ofcourse this bites and Asterisk Developers will not fix this in the ISDN4Linux code inside asterisk. So the patch for this is to rewrite a CallerID of 0 to an empty one using the following in your dialplan:

exten => s/0,1,SetCallerID()
exten => s,1,NoOp

Continue as normal after this.

- Some fixes we made to Asterisk and which I discussed in an earlier entry have been fixed inside Asterisk CVS. The only patch we are currently using is to enable PostgreSQL CDR support. Which is actually adding this:

MODS=cdr_csv.so cdr_manager.so cdr_pgsql.so

to cdr/makefile

- Monitoring Asterisk with Nagios is also something we enabled. We are currently using the following nagios plugin to check it.
   - Nagios Plugin for Monitoring Asterisk
     Original code by Roy Sigurd Karlsbakk, Modified by me to support ISDN4Linux
     Get it here.

Currently we are using an Asterisk PBX server at work to process or incoming calls in queues and to support logging and make things automated. Our setup consists of a Debian Linux server with an HFC-S ISDN PCI card. We use ISDN4Linux to interface the ISDN towards Asterisk.

While we where setting this up we discovered a little problem with this setup. The problem is that when a caller hangs up while there is some automated playback and the next caller calls it will hear some remains of this automated playback. The Asterisk Wiki did not have anything to tell us about this issue. However a Google on 'isdn4linux vbox buffer' did reveil us with the following patch for ISDN4Linux inside the kernel:

Please change the code in hfc_pci.c function hfcpci_clear_fifo_tx :

	bzt->za[MAX_B_FRAMES].z2 = bzt->za[MAX_B_FRAMES].z1;
in
	bzt->za[MAX_B_FRAMES].z2 = bzt->za[MAX_B_FRAMES].z1 -1;


This will solve the problem. Asterisk will function correctly afterwards. 

(more)

Happy new year everyone!

Things that are on the todo list for 2005
- Get Asterisk PBX up and running for Unilogic Networks
- Improve Asterisk's skinny support
- Buy a new computer (probably a cheap DELL)
- Post more useful entries on my log
- Try to write more documents that are useful to people
- Make updaterpsl.pl fully RPSLng compliant.

Thing I've done lately and haven't mentioned
- Got Asterisk running in a test setup
- Upgraded polaris to 1.5Gbyte ram
- Got IPv6 running @home by changing my Cisco SOHO 97 to a Cisco 837
- Fixed up mgetty/new_fax to also print out faxes at work
- Moved to a new office with a new workstation