Wednesday, September 26, 2007
mySQL update
We are planning to move mysql1.nsnoc.com to from 4 to mySQL 5 on Wednesday 10th October (two weeks time).
We are currently testing the platform we are moving to, and will be releasing details on how you can test your dev / staging code against the installation before it goes live shortly.
Any questions on the matter please contact support.
We are currently testing the platform we are moving to, and will be releasing details on how you can test your dev / staging code against the installation before it goes live shortly.
Any questions on the matter please contact support.
Load balancer event
We have experienced a fault with one of our load balancer. Taking it out of service users will have experienced sporadic connectivity to the mail server for a five minute period during the diagnosis and change over. Some web sites will also have been affected for this period.
We failed over to the stand in load balancer as soon as we were able, and will be looking to switch back to the main box as soon as we are confident the issue has been dealt with.
We failed over to the stand in load balancer as soon as we were able, and will be looking to switch back to the main box as soon as we are confident the issue has been dealt with.
Friday, September 21, 2007
Plus.Net and Forwarding
As an advisory to Plus.Net users - it would appear that for some reason they are classing our primary secondary and mail scanning mail clusters as "Dynamic Addresses". While this is not the case, it is not preventing them from currently rejecting mail forwarded to accounts on their servers from our clusters.
Feedback has been along the lines of:
[snip]I'm sorry to hear you have been adversely affected by changed we have made to our email delivery system. The changes were designed to reduce the amount of spam our network receives. This is a generally accepted practice in the industry.
No hosting service should be using dynamic IPs to forward mail, you might want to discuss this with them. I have however, passed on the info you have provided, to see if we can make allowances in this instance.[snip]
We are all behind the less spam option and use many restrictions on our own scanning and non scanning servers - however this misapprehension that we are on a dynamic range will cause users with this ISP who usually use our forwarding service issue until they get their block lists corrected.
We will update this with more information as we get it.
[Update 1605]
The following URL was supplied to us by a client for the PlusNet forum. This surely cannot be the case since this is not entirely practical especially with load balanced clustered systems. We await some more confirmation.
[Update Monday 1052]
Thank you to Bob Pullen from Plus.Net for getting in contact with us to remedy this issue so swiftly. Apparently their list was showing the block in question being largely dynamic, and this has now been corrected.
Feedback has been along the lines of:
[snip]I'm sorry to hear you have been adversely affected by changed we have made to our email delivery system. The changes were designed to reduce the amount of spam our network receives. This is a generally accepted practice in the industry.
No hosting service should be using dynamic IPs to forward mail, you might want to discuss this with them. I have however, passed on the info you have provided, to see if we can make allowances in this instance.[snip]
We are all behind the less spam option and use many restrictions on our own scanning and non scanning servers - however this misapprehension that we are on a dynamic range will cause users with this ISP who usually use our forwarding service issue until they get their block lists corrected.
We will update this with more information as we get it.
[Update 1605]
The following URL was supplied to us by a client for the PlusNet forum. This surely cannot be the case since this is not entirely practical especially with load balanced clustered systems. We await some more confirmation.
[Update Monday 1052]
Thank you to Bob Pullen from Plus.Net for getting in contact with us to remedy this issue so swiftly. Apparently their list was showing the block in question being largely dynamic, and this has now been corrected.
