Source Client Interruptions: Customers on cabhs33 and cabhs34 were experiencing issues with the server component dropping source clients at random times of the day. This was due to the fact that the newer systems had updated operating system kernels that included process prioritization which was incompatible with the shoutcast server executable which caused it to restart internally. We repaired the issue by excluding the shoutcast executable from the process prioritization queue and setting the process priority to maximum value. This issue did not occur on other servers in the SoniXCast cluster as they have an older version of the operating system installed without process prioritization support. However future system updates will include such support, so patches have been rolled out to all servers within the cluster in order to avoid future outages.
Event Scheduling Issues: Related to the previous issue, the transcoders (Auto-DJ) were not properly reloading, meaning that that setting changes and events were also not being reloaded. We’ve repaired this issue as well on all servers.
Development Updates: We’ve replaced the timer which polls services and makes the frontend more responsive for users with an updated timer that uses less resources and resolves the issue of high memory use in the browser.