> hosting > status

Hosting services status

Here are news that will interrest those of you that have a web site, a dns, or a mailbox on my computers.

2012 10 07 • Optical fiber, new IP address

"yod" server, hosting web sites, changed its default network interface to use optical fiber today. The old IPv4 address still works. The new IP addresses are and 2a01:e34:ec10:4ef0::f.

DNS have been updated to point to the new IPs.

Upload speed has increased by a factor 20, and latency to backbone is now about 1 milisecond. :)

2012 09 16 • New connectivity

"yod" server was rebooted several times since a week: It received a new fast hard drive and a new network adapter. This adapter is linked to a optical fiber modem currently under test. Be prepared for a 100/50 Mbps download/upload speed! :)

2011 01 16 • Server upgrade was upgraded today to Debian 6 (squeeze).

2010 07 25 • Electrical failure - hosting several web sites, mailing lists, and some backup servers - was unavailable today between 11:39 and 19:10 GMT, because of an electrical problem on yod.

2010 07 04 • Energy usage check

There was a reboot of yod a few minutes ago, and there will be another one later tonight.
This was necessary to plug a wattmeter in front of the server. The result is not too bad: About 40 watts for yod (serving
Previous hardware "teth" was about 70 W, the one before - named "zain" - was about 140 W.

2010 01 03 • Beware of rm

This afternoon, I erased 3 days work by mistake. The user partition was mounted read-only during the recovery, that took several hours. Apache was unavailable then.

2009 11 30 • zain is dead too

Ouch. The backup system hardware failed today too! The remaining web sites are beeing moved to the new server. It should be back to normal during the night.

2009 11 29 • New box

teth is dead. Please welcome yod! The box is beeing setup rigth now.
15:15 Apache is redirected to zain. Ftp, mail is down. Dhcpd, and domains are done.
15:20 Vpn is done.
16:00 zain is very slow. Move of file system is under way

2009 10 01 • Power supply explosion power supply just exploded, sending pieces of silicate everywhere in the commputer! I have been unable to find a replacement today, all shops beeing closed. I recycled the old tower "zain" to take over. There should not be any differences since the data have been moved physically. Outage was about 2 hours.

2009 07 09 • Electrical failure - hosting several web sites, mailing lists, and some backup servers - was unavailable from Jul 7th 15:30 to Jul 9th 14:05, due to a power failure.

2009 04 13 • Sodium update was updated to debian lenny yesterday afternoon. There was small outages for most of the web sites. Everything should be back to normal, now.

2009 02 01 • Magnesium update was updated to debian lenny today, in the begining of the afternoon.
Mail server was off for 15 minutes. There also was a short outage of 3 websites, as most of them are not localted on that server.

2009 01 27 • Power failure

There was a major power outage in my area.
sodium server was unavailable last sunday from 17:00 to 23:00 local time.

2008 10 19 • Mail server upgrade

The was a major upgrade of my main mail server yesterday night. It was unavailable from 18:51 to 19:26 CEST.
My good old Sendmail is finally gone. Welcome to Exim.
The new setup includes TLS encryption, spam detection, and SPF policies.
If you don't trust cacert, you will see a warning about the new certificate. Then, you can safely accept it.

2008 08 06 • magnesium offline for most of the day!

magnesium account has been suspended by my ISP: Mail system was completly cut off. This service is now resumed. Other services were't affected. Dedibox ISP is not has good as it looks, after all.

2008 07 14 • IPv6 change on sodium, magnesium conectivity lost for 2 hours

magnesium crashed at 6:30 this morning, and was offline until 9:30.

sodium - the main server - changed of IPv6 address. The new IPv6 is 2a01:e35:8a27:50b0::f. DNS hosted here will be updated automatically. Please update your records if you have your own.

2008 03 13 • IPv6 on magnesium too is now IPv6 enabled. Mail server, and secondary DNS are now available using that protocol. The kernel has been upgraded to 2.6.22 and a reboot caused the machine to be unavailable from a minute.

The new IP address is 2a01:e0b:1:38:240:63ff:fee7:b1b6

2008 02 05 • Offline 5 minutes

Teth's old hard drive has been removed, and everything now runs on the one installed last june. The machine was shutdown for a few minutes.

2008 01 28 • Feed

An experimental feed is now available for this page.

2008 01 08 • IPv6

teth server - aka sodium, the main server - is now available on IPv6 at 2a01:5d8:58a2:750b::f

2007 08 23 • Nuclear free site

Since august 1st, my computers are powered by Énercoop, a cooperative company providing electricity from 100% renewable sources.

2007 08 16 • Modem gone for a few hours

I lent my modem to a coworker, so he could get his freephonie certificate. This was supposed to shut down my connection for an hour or so, but I went out of my flat leaving the keys inside, so that it took longer than I expected. Teth server has been offline a while, so access to web sites, ftp services and some of the mailing lists has been impossible during that period. Others services such as DNS and mail are handled by the backup server.

2007 06 28 • More tuning

I'm updating my network configuration here, and it will lead to several small connection cuts today, less than a minute.

2007 06 26 • Statistics

Unbelievable, after all these years, I finally installed a web usage stastics generator! Go now to to see yours!

2007 06 25 • Hard drive upgrade

I upgraded teth server hard drive. System has been unavailable for 2 hours tonight. Everything should be fine now.

2007 02 16 • Modem crash

My modem went offline for no known reason. A eletrical restart was necessary. The web site and ftp server has been unavailable for about 10 hours today. If changed some settings and hope this won't occur too often.

2007 02 13 • Modem crash

My modem went offline for no known reason. A eletrical restart was necessary. The web site and ftp server has been unavailable for about 8 hours today.

2007 01 20 • Servers IP changes

I plan to remove It will be replaced by at first. Pratically, I'm switching sodium and neon IP adresses, on zain and teth servers.

I installed a 4th level backup DNS service on zain server. That server' public address was and is now becoming That service will be maintained a while, ensuring there's no cache problem.

HTTP port at zain is beeing redirected to teth, so that there will be no http service outage.

I installed a DNS service at zain (now neon) and updated all DNS details so that sodium is now officially in NS lists. Furthermore, sodium is now the Source Of Authority for domains I handle directly.

The swap occured around 13:00 GMT+1 and downtime was less that a minute.

Please use your domain name to access your ftp. There's no ftp service on neon, use sodium.

2007 01 07 • Mail service move

The primary MX has been switched to People running pop3s have been mailed (or phoned ;) about the changes in their set up. Only mailman is still running on neon.

2007 01 01 • New server & DNS updates

I finally rented a dedicated host in a telehouse. Its name is The bandwidth is great, about 100Mb/s. That's a great improvement.

A DNS server is allready running there. All domains that I handle directly have been updated so that their first DNS server is magnesium. People whom I handle their backup DNS only have been mailed, with technical data.

Happy new year to you all.

2006 09 03 • Repartitionning

I reparition main server tonight and did a lot of clean up. DNS and Mail backup took over. However, both web and ftp services were unavailable for a while.
I also installed a new mail server: postfix. New certificates were generated, so don't be surprised if you get a warning. Please tell me if you encounter any problem.

2006 08 19 • WiFi

There as been a few reboots tonight for hardware updates.

2006 08 17 • SORBS

I finally installed SORBS this week. It's a mail filtering system that blocks SMTP access from servers known to have a dynamic IP address or that usually spams. Fell free to contact me if you have trouble beeing mailed to on your hosted mail addesses.

2006 03 05 • Disc full

Primary partition has been full for one day, because I forgot to deactivate the logging of debugging messages. Mail delivery has been delayed during that period.

2005 10 05 • Offline for 9 hours

The power supply was accidently switched off between 1:36 and 10:39 (CEST).

2004 10 05 • Network adapter failed. backup server used for 2 hours

The main box has been droping packets since a few days. The network adapter finally failed (hardware) at 2 o'clock this morning (CEST). The main server was offline for 2 hours. The backup server took over all services but http during that period.

2004 09 09 • PHP

The main server has been updated again today at 15:50 GMT+1. PHP is now available on all web sites. The server has been rebooted and was unavailable for a very short while.

2004 09 05 • Updates

The main server has been updated today at 12:45 GMT+1: A new kernel and a new version of apache have been installed. The server has been rebooted, and was unavailable for a very short while.

2004 03 24 • Apache unavailable for 3 hours

I updated the kernel on the primary IP and something went wrong with apache configuration. The http deamon displayed no documents for 3 hours today, starting at 15:35 GMT+1.

2003 12 19 • pop3s unavailable

I updated the firewall yesterday and missed the secure pop3 port that has been unavailable for a whole day. Sorry about that.

2003 12 18 • Primary box offline

At 01:30 CET, the DSLAM from my operator went offline. All the neighbourhood was offline for about 12 hours.

2003 12 04 • Main box was restarted

At 17:30 CET, teth, the box handling the primary IP connection was disconnected several times for about 10 minutes in order to upgrade several components, including the adsl modem.

2003 10 05 • SSL upgraded

Pop3s service and smtp tsl option are back.

2003 10 02 • pop3s service interruption

Pop3s service is down for security reasons. Plaintext pop3 (110) is temporary available. Everyone with a mailbox here has been mailed about this, both locally and on remote systems.
Additionally, SMTP TLS option has been deactivated.

2003 10 02 • Backup line info

I gonna change of ISP for the backup line too ( The backup IP address will change. Today, I deactivated one of the backup DNS - leaving only two dns servers online. I hope the new connection will be up this week end or early next week.
The backup server is now longer forwarding http requests to the new primary IP (see 2003 09 07 announcement).

2003 10 01 • Primary server offline during almost 5 hours

The main IP stop responding at 15:30. My ISP, Free, decided to install extra equipment in its Voltaire site. As usual, there has been no warning whatsoever, and thouthands of people got disconnected for many hours. *sigh*
The connection was restored at 20:10 CEST.

2003 09 16 • Primary server went offline an hour

In the end of the afternoon, the main server - the one hosting web sites - got disconnected for an unknown period of time, probably less than an hour. The cause is a faulty ethernet wire.

2003 09 07 • Primary & backup lines switch

I'm switching the primary and the backup services betweeen and fluor is 3 times faster and will become the primary IP.
The first step is the DNS and mail services. Everyone have been mailed about what they need to do. The old services will stay online until I got confirmation everything is ok.
17:22 CEST: The filesystem is being moved to the new computer. FTP IS DOWN.
18:04 CEST: The new server is up and running. Please stand by for more information...
18:34 CEST: The switch is complete. The FTP is online and can be reached on Please update your dns to use instead of carbon. For a while, carbon is internally forwaring http requests so that the service will work on both adresses.

2003 09 06 • New box "teth"

A new computer called teth is now online. It's a 2.4GHz Pentium 4 with 1GB memory and a 120 GB hard drive. It is now handling IP, the backup DNS and mail servers. It will become the main server next week.

2003 09 05 • Backup line info

The backup line,, has been stable for a week and then began again droping packets from time to time. Today, a France Telecom guy came and fixed "an error in wiring". The modem stoped complaining about noise on the line and about ATM receiver overflows. Let's hope the problem is fixed! :)

2003 08 10 • Reboot

The main box kernel has been upgraded to version 2.4.21. All services have been unavailable for a minute, at 17:23 CEST.

2003 08 10 • Backup line info

The backup line is now using a new provider, Free. I've been using it for over a week now. The new IP address as been named is down and will not be used anymore. The output rate of the new line is better than expected: over 42K output. However, it's still unstable: It requiere a modem reboot every day or so. I'll not swich primary and backup lines until this problem is solved, of course.

2003 07 31 • Primary connection lost for 4 hours. No more backup line.

I'm upgrading my adsl connection. The backup line is going from 64K/16K to 128K/32K. The change occured at 10:32 GMT+2. However, a cable was not like I tough and it was the primary line that went down. It was not restored until until 14:45 GMT+2.
My new DSL provider requiere a lot of special settings, so that it will take a while to get that backup line online.

2003 07 14 • PPP failure

The connection on the primary server failed at 6:56 GMT+2. The associated ADSL modem just didn't want to communicate. The system went into a loop eating all resources. The system was hard reboot at 14:12 GMT+2 and the modem at 14:20.

2003 07 12 • HTTPD update

The web server was shut down for 7 minutes in order to update it, starting at 16:36 GMT+2.

2003 07 09 • Passive FTP

Firewall has been modified so that passive ftp connections now are allowed. Enjoy! :)

2003 07 08 • Shutdown for 15 minutes

This afternoon, I had to power off both primary and backup servers in order to rewire my hot water system. Sorry about that.

2003 06 25 • SMTP demons now are SSL enabled

Both SMTP servers now support the STARTTLS command, so that mails are encrypted during the transfers between servers. However, the remote server has to support that extension, a few of them do. Mails to/from low security server are transfer in plain text, as usual.

2003 06 18 • Server updates

The servers are now using a new 2.4.20 linux kernel. The outage was less than a minute.

2003 06 16 • 32K output bandwith

For the last 24 hours, EQL driver was setup on the main box: The two lines were agregated so that the output bandwith reached 32K. It was temporary and the services now are back to usual configuration.

2003 06 06 • Waw is back

Waw, the main web is back online! The crash was caused by a defective video card! The installation will be tuned in the next days, and services may be interrupted briefly during that period.

2003 06 04 • Hard disk crash

Waw, the computer hosting webs crashed yesterday around 15:00. All data was lost
Zain, the backup mail/dns server took over http service today. A backup from april 2nd was restored on it. A few updates were recovered from the lost+found directory in the failed hard drive. All residents were warned: Please do check your web site completly!

2003 04 03 • Perturbations

Systems have been shutdown yesterday and today for some short whiles. Startup scripts were tested (more to do there). The power supply failed in my block for a few minutes early this morning. And finally I tested my new Ground supply around 9:00 GMT, and forgot to bring back the http deamon afterwards.

2003 02 11 • Electrical tests went wrong

I tested today my brand new electrical board: Things didn't exactly went as planed. All the systems have been down for several hours, until EDF had the building general system fixed. Ops! :]

2003 02 03 • I moved

The was no significant outage. :)

2003 01 02 • Move scheduled

All the servers will move physically in about two weeks.
A backup http server has been setup today. The network reconfiguration caused a few small disconnections yesterday and today. During that time, I also added a new 80GB hard drive in the main server. It should take a while before there is any space problem. ;)
The http service now has a backup, so that the service shouldn't be affected much.

2002 12 30 • DNS server update

The DNS server has been updated at 01:00 GMT+1. The backup server was used for a few minutes. There was no outage.

2002 12 12 • Connectivity lost for 5 hours

Internet connectivity was lost between 01:42 and 07:11 GMT+1. Log files show a huge number of keyboard errors then a reboot of the gateway. I have no other evidence, but I strongly suspect my cat.

2002 09 29 • Connectivity lost for a few minutes

Connectivity was disrupted sereval times yesterday night, because big red glowing buttons - that is general power shutdown for the whole computer room - are very very attractive for a two year old kid.

2002 09 24 • Connectivity lost for 6 hours

Connectivity was lost between 06:00 and 12:00 GMT+2. I guess the connect script need another update

2002 09 05 • New net trouble ahead?

I got a new inhabitant here. I saw him trying to log on my Linux box. He tried to sabotage my network too. Here's the proof:

I hope there will be no real disruption though.

2002 08 26 • Connectivity lost for 4 hours

Connectivity was lost between 07:00 and 11:00 GMT+2, because of an error in the connect script on the new box. This has been fixed and should not occur again.

2002 08 25 • System update

The whole system has been moved to a Linux box around 18:00 GMT+2. HTTP server is exactly the same. The changes are significative in the FTP server however.
Your initial folder is /www. That is where your www documents are. You can go back one level in the folder tree if you wanna see the logs or change your local HTTP configuration.
The local file system default end line sequence is now <CR> instead of the m$ style <CR><LF>. Just remember to transfer your file in BINARY mode.

2002 08 18 • Apache upgrade

The HTTP server has been upgraded around 16:00 GMT+2. File system maintenance requiered a 20 minute FTP shut down. Your HTTP configuration files are still valid with new version: Apache 2.

2002 07 29 • Gateway upgrade complete

Connectivity is back on the main router, that has been upgraded.

2002 07 27 • DNS reverse update

I finnaly got a proper reverse for the main IP This shouldn't have any consequences on hosting.

2002 07 16 • Gateway upgrade starts

I switched the network on an alternate router around 16:00 GMT+2. The FTP service has been shut down for about an hour.

2002 07 01 • Storm

A power supply failure caused a few minutes downtime around 15:20 (GMT+2).

2002 06 14 • Power supply failure

A global power suppply failure in Paris 19th district around 23:40 (GMT+2) caused a 5 minutes downtime.

2002 05 02 • DSLAM change

Since today, I am using a new DSLAM.
May 9th update: Since then, the daily disconnection never lasted more than 3 seconds. :)

2002 05 01 • No more backup IP adress

The old IP adress is no longer working.

2002 04 30 • Network maintenance sheduled for May 2nd

I finally sent the some low level PPPoe connection logs to France Télécom. These show the BAS saying Internal Error: "No system resources"! After a few exchanges, the problem has been localised and they'll route my connection via another DSLAM starting next thrudsay.
You can expect some disruption for a few hour that day. Sorry for the inconvenience.

2002 03 15 • Connection failure

I've been away from my system for 5 days. It looks like the switch power suply failed the first day I was out. Sorry for the inconvenience.

2002 01 12 • IP change

I changed my primary IP adress from to The http port of the former IP adress is being redirected to the new IP adress so that there is no service outage. Please use the new address for the other services immediatly. The web redirection will probably work for a few monthes, but you should try to update your dns asap.

2001 12 06 • 3 days offline

I had to leave the server without supervision for a week and of course, Murphy's law was enforced. The connexion went down December 3rd and wasn't restored until today. Sorry for the inconvenience.

2001 10 24 • All system are working

All services have been restored and maintenance is complete.

2001 10 24 • Network Address Translation setup complete

A better setup of the internal network will now enable you to see where HTTP connections come from. Logs will sure make more sense that way.
FTP ans SOCKS services are still down.
Yesterday, I successfully switched my old modem to use pppoe, so that the gateway now has several IP adresses. Of course, the plan is to use brandwith load balancing. :)

2001 10 23 • Problems with apache virtual hosting fixed

The IP adress change of the apache server made yesterday was not immedialy reflectled on apache configuration, so that all virtual hosting matches failed. As a result, every one trying to connect on an http site here would see content!
FTP and SOCKS services are still down.

2001 10 22 • New gateway

The installation of a new Linux gateway is under way. It is now activated. However, all the services haven't been brought back yet. For exemple, you can't use FTP rigth now. Stay tuned for updates...

2001 10 19 • Power supply failure

Connecticity to the internet was lost between 13:00 and 21:00. Power supply was shutdown in my building. I had a shock listening to the pure silence when I came back home!

2001 10 08 • Connectivity lost for 10 hours

Connecticity to the internet was lost between 9:00 and 19:00. The auxilairy host that handles conectivity mantenance was restarted the 7th but the connectivity tools wasn't restarted so that nothing took measures when France Telecom shut down the ADSL line at 9 o'clock. Sorry for the inconvenience.

2001 09 22 • Connectivity lost for 3 hours

Connecticity to the internet was lost between 9 o'clock and noon. The auxilairy host that handles conectivity mantenance had been shutdown for the nigth so that nothing took measures when France Telecom shut down the ADSL line at 9 o'clock. Sorry for the inconvenience.