5,336
edits
(Created page with "Notes on high-availability: keeping your services running even when some underlying resources need to go offline, e.g. for maintenance or due to failure.<br> Note that I do not run any actual highly available services so I have no experience here. ==Kubernetes== ===Load Balancing=== See [https://metallb.universe.tf/concepts/layer2/ Metallb layer2 mode].<br> Here you have two instance of metallb running. In layer 2 mode, metallb gets an IP address on the network, separat...") |
|||
Line 8: | Line 8: | ||
In layer 2 mode, metallb gets an IP address on the network, separate from the IP address running on the node. | In layer 2 mode, metallb gets an IP address on the network, separate from the IP address running on the node. | ||
When one instance fails, the other instance takes over that IP address so the routers still know where to forward packets. | When one instance fails, the other instance takes over that IP address so the routers still know where to forward packets. | ||
====BGP==== | |||
===Storage=== | ===Storage=== |