We need to update both Server SGC & Atlantis MSSQL Instances to the latest version.

This will be done between 00:30 and 01:00 on Sunday 24th Feb

This will take around 10mins and shouldn’t require a reboot.

Control Panel, Primary SMTP, POP3, Secondary SMTP, Dhosting Helpdesk, Dhosting Domain Reg and any website that uses a MSSQL DB will be offline during the update.

Update 24/02/2008 00:55 - This work is now complete; Server SGC MSSQL Instance is now at Build 9.00.3200.00. We will be monitoring the server for any issues over the next 24/48 hours.

Update 24/02/2008 01:21 - This work is now complete; Server Atlantis MSSQL Instance is now at Build 9.00.3200.00. We will be monitoring the server for any issues over the next 24/48 hours.

On Friday 22th Feb 2008 between 22:30 and 23:30 we will be upgrading the memory on Server SGC from 2Gb to 4Gb.

This work will take approx 5-10mins while we shutdown the server, add the ram then boot it up again.

During this maintenance period websites hosted on Web1, Primary DNS, Secondary Mail and Helm Control Panel will be offline.

Primary Mail, Secondary DNS and all websites hosted on Web2 will be unaffected by the work

We appolgise for any inconvenience this may cause.

Between 5:05pm and 5:18pm today users will have noticed high levels of packet loss to our servers both in London and Kent.

This was due to a DoS attack directed at one of our network provider’s customers.

The DoS attack was filtered and network traffic returned to normal.

Update 17:58 - Our network provider have informed us the network upgrades which couldn’t be completed last night will be done tonight. The work had been re-scheduled for Friday but in light of the attack detailed above they want to get the extra capacity live a.s.a.p. If the upgrade had been completed last night its more than likely today’s issue wouldn’t have been noticed.

The network upgrade will be performed between 20:15 and 20:45 this evening. Customers will notice up to 2 network blips at around 60 seconds each while the network ports re-negotiate.

This work only affects our servers in Kent including Server Atlantis and all Managed Servers customers

Update 20:54 - Network Upgrade is now complete, total downtime 35 seconds.

At around 3:45pm today Server SGC started to show high levels of packet loss. Access via HTTP was extremely slow with some users getting time outs.

Initially the server would not let us access it using RDP, so we remotely restarted various services including IIS and MySQL 4, which we suspected were causing the issue.

After gaining RDP access we restarted various services with no effect, there was still a performance issue but websites were accessible unlike at 3:45pm.

Windows was showing no errors or high CPU / Memory / Disk Usage. Only one service had not been restarted which was MSSQL 2005.

We decided to give MSSQL 2005 a restart and this seems to have cleared the performance issue with website access now being instant. There are again no errors in the MSSQL Logs, no sign of large amounts of connections or a script looping.

This server hasn’t been rebooted in over 60 days which isn’t a problem as we have web servers that haven’t been rebooted in 4-6 months. But it seems MSSQL was struggling.

We will be monitoring the situation very closely over the next 24hrs to make sure it doesn’t reoccur again.

A reboot will be scheduled for both web1 and web2 at the weekend, we also intend to double web1’s current memory from 2Gb to 4Gb when we reboot it.

BT will be making a visit to fix our main phone line this morning at around 9:45am. This work may take up to 2hrs.

We will be redirecting calls but due to the main line and ADSL both being offline it may take slightly longer to answer calls.

This work should finally fix a noise fault on our line.

Update 12:00 - Work has been completed on the main line.

Our network provider will be upgrading their main trunk into Kent Science Park to add an additional 800Mbps capacity. The upgrade will take place between 19:45 and 20:15 on Tuesday 12th February 2008. During this small window customers are likely to experience between 30 and 90 seconds loss of service while the ports re-negotiate at higher speeds.

We apologise for any inconvenience this may cause to customers, however we must also stress the upgrade is crucial to ensure we provide uncongested high quality bandwidth to clients located in Kent.

This will affect all websites hosted on web2.gbdns.net and primary mail services, including POP3 / IMAP and SMTP. Our backup mail servers will catch all mail which fails to be delivered to the primary mail server.

Websites hosted on web1.gbdns.net in Telehouse East will be unaffected by this work

Update 13/02/2008 01:30 - This upgrade was not attempted due to a fault being found in a piece of Fiber Optic Hardware. This will be re-scheduled and we will post an update when we have a date and time.