Skip to content

Player FM Status

Just the facts ma'am about upgrades and downtime

There has been a syncing issue over the past few days which caused some feeds to not be updated correctly.

The main cause was over-excited logging. After the recent infrastructure upgrade, there was a problem that caused logs to not be cleared correctly, impacting on both performance and memory usage. This has been fixed now and should also avert website and API slowdown. I’ve also simulated the cleanup process to ensure it won’t happen again.

The main feed-processing servers have also now been turbo-charged, thanks to Linode’s recent and delightful mega-upgrade. It doubles RAM (which was already doubled a year ago!), brings faster processing by way of SSDs and processor upgrades, and improves network speed. Altogether a big win.

Having said that, I’m still aware there may be another issue with some feeds being marked as stale and not reverting from that. I’m still monitoring carefully and meanwhile I’ve also been developing an open source library which will ultimately make the whole process more robust and performant. UPDATE: Identified and fixed the issue. The regular re-fetch cycle wasn’t happening due to a software bug that slipped through (yes, there’s now test coverage on it). The reason it wasn’t detected is feeds were still being updated upon receiving push notifications, so it was only those where a notification was missed that were never re-tried.