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
We are currently using nginx as a deployment. In front of our ingress controller we have a F5 Big IP AWAF Instance which is doing the "Loadbalancing". Due to the current implementation there is no real Loadbalancing since 1 Backend and ServiceType Loadbalancer in the cluster. We want to swtich to NodePort version of the service. So we can configure on Big IP a backend pool with all our nodes included. This being said and tested works. Due to lifecycle and troubleshooting reason we are often required to recreate a node (delete and readd one) or restart a node. For this we do the normal drain and then shutdown the node.
In case of the NodePort and NGINX deployed as DaemonSet all still open TCP Session to the node which is powered off are stale and waiting for their Timeout TCP FIN Request. This can cause outages which are ~ 1min. I want to prevent this. Do you have any recommandation for the setup how this can be achieved. I'm pretty sure I'm not the only one struggling with this
+-----------------------+
| Client |
| (HTTP Requests |
+-----------------------+
|
v
+-----------------------+
| BIG IP F5 |
+-----------------------+
| |
v v
+---------+ +----------+
| Node 1 | | Node 2 |
| (dies) | | (healthy)|
| No FIN | | |
+---------+ +----------+
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
We are currently using nginx as a deployment. In front of our ingress controller we have a F5 Big IP AWAF Instance which is doing the "Loadbalancing". Due to the current implementation there is no real Loadbalancing since 1 Backend and ServiceType Loadbalancer in the cluster. We want to swtich to NodePort version of the service. So we can configure on Big IP a backend pool with all our nodes included. This being said and tested works. Due to lifecycle and troubleshooting reason we are often required to recreate a node (delete and readd one) or restart a node. For this we do the normal drain and then shutdown the node.
In case of the NodePort and NGINX deployed as DaemonSet all still open TCP Session to the node which is powered off are stale and waiting for their Timeout TCP FIN Request. This can cause outages which are ~ 1min. I want to prevent this. Do you have any recommandation for the setup how this can be achieved. I'm pretty sure I'm not the only one struggling with this
Beta Was this translation helpful? Give feedback.
All reactions