Redundancy behavior if one of instances down

[ilfarpro]

I’ve read documentation segment about redundancy, but still have few questions.

1. What happens if my instance video is replicated on other instance but that instance is down? What would I see in my original instance? Bufferisation issues? Failed http queries to second instance that is down?

2. What happens if my instance goes down, but it’s videos are replicated on other instance? Will the videos from my inaccessible instance play on that instance without issues?

[Chocobozzz]

@ilfarpro

What happens if my instance video is replicated on other instance but that instance is down? What would I see in my original instance? Bufferisation issues? Failed http queries to second instance that is down?

After HTTP request failure, the client will remove the instance that is down from its sources and will retry the request on another instance.

@ilfarpro

What happens if my instance goes down, but it’s videos are replicated on other instance? Will the videos from my inaccessible instance play on that instance without issues?

If your instance is down the video cannot be played. PeerTube redundancy is not a « failover » system. Some files are still fetched from the origin instance (m3u8 playlists for example)

[ilfarpro]

Hi, @Chocobozzz ! Appreciate your answers.

@Chocobozzz

client will remove the instance that is down from its sources

1.How fast it happens? Will end user experience bufferisation?
2. When client (browser) will retry HTTP requests to instance that is down to verify if it’s available again? Maybe when I update page or open another video?

@Chocobozzz

PeerTube redundancy is not a « failover » system

Is there any chance that in future it will be like failover?