VPS8 Issue Resolved and Full Information
-
Friday, 20th December, 2013
-
09:12am
This is some information on why VPS8 has had a number of reboots in the past 10 days after the server became unresponsive. We discovered the server was reporting out of memory errors and on a number of occasions we were seeing packet loss. Packet loss appeared to happen when the Node Based backup was running that copies the data to a different server in a different location and investigations lead us to discover the port speed of the server was set to a very low 10mbps. This clearly was an error and we have corrected that with help from our data centre partners who ensured the switch port speed was at full capacity.
Unfortunately as this server had a long uptime the first reboot required a file system check that took a few hours last week.
We are pretty sure the out of memory errors were caused by the servcer traffic bottlenecking a direct cause of the low port speed but just in case and purely as a precaution we doubled the RAM in this node. This is why this node was switched off for 30 minutes at 3am this morning.
We are very confident this will fully resolve any issues from the past 7 days and whilst most outages were just a few minutes we are extremely sorry that this node has not lived up to the high levels of service we always strive to provide.
