LightSpeedTaco

Following a recent 3 day stint of cable making, the network is now running on shielded cat 6 networking cables that are all hand made and tested by myself. They are cut to length to minimize airflow interference. With this upgrade the potential for introduce LCAP which is link aggregation and could allow up to 4 gb/s transfer rate between the NAS and the servers thus speeding up overall speeds. This is still pending as an appropriate switch will need to be integrated. Next, the servers are crammed onto my desk and have my keyboard and mouse ontop of them. This isn’t ideal for my comfort and heat distribution/airflow for the servers themselves. Therefore some physical rearrangement will be necessary and this is the cause of the downtime. We expect to be down from 2AM EST to 8 AM EST. Keep an eye on the Nobreaks.ca twitter for details and status updates. Regards, The LightSpeed Team

Hello everyone, So the journey has lead us to some large scale processing power. Currently the data is stored and collected on the 20 bay (NAS + Expansion) storage solution with a potential of 120 TBs of protected RAID 6 storage and is now processed on a cluster of 3 Dell PowerEdge R610’s each with 9 146 GB 10K RPM 6GB/s SAS hard drives in a RAID 6 array. Each server has two Intel Xeon processors and 48 gigabytes of RAM running ESXi 6 and virtual Ubuntu 16.04 systems. Rather than having multiple plex servers I have implemented a script made by a fellow Plex user over at the Plex forums that is written in Python and in real time, load balances the transcoding of media being requested on Plex. This way, the burden of transcoding is spread out among 6 physical CPUs for the smoothest playback experience. Over on the “Is the Server Up” status page you’ll find a new link to the “Cluster Load” page which details the CPU load of each node, the number of processes running, uptime, connectivity, RAM utilization and disk usage and is updated every 5 minutes. Below are some photos of the new… Read More »From NAS to Cluster

Hey everyone, In the next 2 weeks or so I’ll be purchasing a second Dell PowerEdge R610 which will be used for both real-time load balancing for Plex as well as failover for the entire VM the Plex Media Server is hosted on. This way, if one server or VM goes down it will be shifted to the other working server. The networking is already redundant, as is the storage and soon the entire system will be. In regards to the real time load balancing, this leaves room for an unlimited amount of servers which can carry the load of transcoding and serving up media meaning a higher quality experience for you all. Regards, Team LightSpeed

Hey everyone, You may have noticed some inconsistency with the uptime of the Plex Media Server recently. This is due to some on the fly fine tuning going on involving the hardware assigned to the server’s VM instance. The NAS had 4 ethernet connections which were bonded via the QNAP control panel where the new server which also has 4 ethernet connections is bonded via ifconfig and this is where I found some issues with overhead and issues distributing load efficiently. I was also attempting to implement remote transcoding which would allow multiple physical servers to work on transcoding  together however with the recent dawn of v1.0 of Plex Media Server has thrown a wrench into the script which I haven’t been able to circumvent. For now I’m going to say things are stable and quit messing with them. Regards, Team LightSpeed

Hey everyone, Not much was said about this because I didn’t necessarily expect it to work the way it did but here we are. Whats happened is that I’ve invested in a 1U Dell PowerEdge R610 with dual quad core Xeon processors, 48 gigs of DDR3 RAM and 600 gigs of RAID 6 redundant SAS (10k RPM) storage. What I’ve done is give this new server running ESXi 6 access to the NAS where all the content for Plex is stored, gathered and usually served from and setup a Ubuntu 16.04 64 bit VM to act as the new transcoding server thus taking the weight off the NAS. Now, it’s not that the NAS was not capable of keeping up but rather this monster server could do it’s job twice as well. With a total of 8 cores the CPU intensive process of transcoding media for users would be much better handled and leave the NAS to do what it does best. Store data. While many of the services I.E the PlexRequests server run on the NAS, the newly reopened minecraft server and the transcoding (Plex Media Server itself) aspect of things now has its own dedicated hardware. Please let… Read More »The Great Cutover