Tuesday, July 28, 2009
Cluster 150 - 5.15pm
It would appear that a customer wordpress site has caused the web cluster to overload. We're currently waiting on the load to fall on the web servers to confirm the cause and to restore service.
We expect the servers in cluster 150 to have recovered within the next 5 to 10 minutes.
An update will follow shortly.
UPDATE - 5.25pm
Services have been restored and will be monitored over the course of the evening. The service restart prevented us confirming the exact process details that had caused the server load to climb.
We expect the servers in cluster 150 to have recovered within the next 5 to 10 minutes.
An update will follow shortly.
UPDATE - 5.25pm
Services have been restored and will be monitored over the course of the evening. The service restart prevented us confirming the exact process details that had caused the server load to climb.
Monday, July 20, 2009
"delayed emails" / duplicates.
Friday saw a scheduled file system check of a mail cluster member. It would appear that a complication with files not being deleted properly after being cleared from the spool queue had arisen, and as a result on recovering these files to a state in which they could be deleted - the mail server diligently sent them on their way. The resulting phanton resending of emails - in some cases up to a month old.
As these mails had been marked as deleted, these mails are DUPLICATES, and these and can safely be disregarded.
We apologise for the confusion this may have caused.
As these mails had been marked as deleted, these mails are DUPLICATES, and these and can safely be disregarded.
We apologise for the confusion this may have caused.
