You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
1 idleTimeoutExpired
2 channel closed by remote end
3 channel TTL expired
But there is one special case making channel not usable. We use AHC to connect to a service behind AWS ALB. When a new deployment happens, the DNS of service will be changed targeting new deployed vesion . But for a period of time(say 1 hour), old version and new version services co-exist. During the hour, the channel of old version service can still be used . But that's not as expected.
Could we update the DefaultChannelPool.IdleChannelDetector to add check
if channel's remoteAddess is not in list of DNS resolve result , we will mark this channel as expired.
At the moment a channel will expire when either
1 idleTimeoutExpired
2 channel closed by remote end
3 channel TTL expired
But there is one special case making channel not usable. We use AHC to connect to a service behind AWS ALB. When a new deployment happens, the DNS of service will be changed targeting new deployed vesion . But for a period of time(say 1 hour), old version and new version services co-exist. During the hour, the channel of old version service can still be used . But that's not as expected.
Could we update the DefaultChannelPool.IdleChannelDetector to add check
if channel's remoteAddess is not in list of DNS resolve result , we will mark this channel as expired.
Happy to raise a PR for this.
code in DefaultChannelPool.IdleChannelDetector
The text was updated successfully, but these errors were encountered: