Our activation system (i.e. e-mail confirmation) has been off for a few days – it’s been fixed up now. There was quite a lot of emails that were unsent. So we pruned the queue, all activation emails that were requested were pruned, all private PM notification emails were pruned, all topic reply notification emails were pruned. This decreased the queue by about 30K to 15K atm. The emailer will begin to send emails as of now.

The reason why the emailer system broke down is because the recent DoS to the network has caused MySQL to notice failed connections from our mail server and denied it access.

We have fixed the erroneous script that caused some legitimate users to be blocked from our site.

The script used to block DDOS could have erroneously blocked legitimate users from accessing the site if they had accessed the site several times when the swarm of bots caused the DDOS.

Members with dyamic IP who find the site unreachable are requested to renew their IP addresses to access the site. Members with a static IP are requested to use proxies to access the site temporarily.

The blocklist will be fixed soon.

We’re experiencing a DDOS attack directed at our server. We’re working on a script to tackle this issue. We regret the inconvenience caused.

Apparently the main routing into our datacenter was cut unexpectedly. We went back up on a back up link and the main routing is being restored as we speak.

We’re facing thermal issues and so we’ll be replacing our server fans and relocating our servers. This will cause downtime. Please be patient. Thanks for your understanding.

Update:

Server has been shifted to a congenial environment.

It appears the error You cannot make another post so soon after your last; please try again in a short while happening to some of our members who posted a reply or PMed someone when our server’s times were off. It currently looks the servers were out of sync by about a year. The reason why our times were off by a year in that short period was due to the kernel upgrade.

FYI, our servers are synced with primary NTP time servers every 15-30 minutes to ensure all our servers are in sync with each other and are in sync with the actual time.

This bug will be fixed shortly.

Finally after a day, it looks like the kernel problem that’s prevented our upgrade to the latest version has been fixed. Now I feel stupid, it was due to overlooking a configuration setting.

It seems like the kernel upgrade will take longer than expected due to complications.

We’re currently building the latest version of the kernel for our Xeon database server. This should result in few minutes of downtime because we will need to boot into the new kernel. Because of the restart, we may lose some topic views counts. This is because the topic views is stored in memory first before it is written to the database to decrease load on the server. But we will try and minimize the view count lost.

It may be a few hours before we actually restart the server because it takes a while to build the kernel.

PS Other site Admins, heads up, PHP 5.3.0 is out and it is expected to improve performance on PHP in general.