Welcome to Platypus Media. We have 650 clients hosted on Brisbane and Sydney servers and use this website on a Melbourne server to announce any news about Sydney and Brisbane servers.
Brisbane Linux server Platypus:
22 Sept 2016: Brisbane network issue –
The cause of the problem today was the SAS SAN for the cloud went offline
The cause of this was the failure of the power board which had 2 of the power supplies to the SAN failing
In each cabinet in the Data Centre we have 2 separate power feeds an A and a B so that if one power feed fails we have the second one still live
Our equipment therefore typically has dual power supplies with one plug into the A feed power board and one into the B feed power board
Because the SAN has 3 power supplies it required 2 to be in one of the power boards and the 3rd in the other
When the power board with 2 of the SAN plugs in it failed the SAN could not operate on the 3rd alone. Something we were not aware of and I personally feel is a design issue.
I say this because I am not aware of any local DC which has 3 separate power feeds per cabinet to allow the 3 plugs to be on separate feeds
The result was it went offline, the plugs had to be relocated and the SAN brought back up. Once the SAN was online we have to restart each server. The system would go into overload if it tried to start them all at once so it queues them. Hence the staggered progress of them coming back online
With our impending upgrade the SANs will become redundant and we will be using Hypervisors with Integrated storage
What this means that instead of storage being on a SAN with RAID the storage will be on the actual Hypervisors which will all be interconnected and each drive will be mirrored to another Hypervisor
This means that if a Hypervisor fails all servers in it will start up on another Hypervisor that has the mirror of its disk
This is the latest Cloud technology and removes the potential for a repeat of today’s incident
Caches: MemcacheD is installed for this server, but requires a plugin for it to work. WP Super Cache is also recommended.
Use http://tools.pingdom.com/ to test speed before and after but make both tests as close in time as possible as the time of day or night as well as where the test is from, make a difference to loading speeds.
We also recommend the www.cloudflare.com cache for all sites with a high volume traffic.
Daily backups are available to website owners in cpanel – Files – R1Soft. These may be downloaded or used to restore all or any main folder.
Use the cpanel login and password for this. This is in addition to cpanel Backup Wizard.
Brisbane Windows server Platypus: Last down 8 March 2016 6:15pm.
Hyper-V, codenamed Viridian formerly known as Windows Server Virtualization stopped. Techs migrated all accounts to other HV’s after which the server was rebooted. Emails still worked.
Sydney Linux server Hiruko : Still using PHP version 5.4.31 and MySQL version 5.5.40-cll – any wordpress-woocommerce clients should call us
Last down 18 April 2015 – cpanel and FTP stopped working although websites were still working. Fixed next day.