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
With the issues I ran into in #392 I considered going back to VIP mode given the HA provided by keepalived (when it doesn't keep crashing!), might be useful.
However in thinking about doing this I realized that there may be conflicts with virtual_router_id with multiple instances/clusters on the same network.
This, among other customizations, I believe the template used for the keepalived.tmpl should be made available as a config-map in the cluster so the user can adjust as needed.
Also concerned about the flushing of IPVS tables when keepalived starts up, so access to the daemonset deployment template would also be useful, though I realize this could be edited after the cluster is deployed, with openelb managing that it feels there may be another mechanism needed to edit, or just leave deployment of that outside of openelb, which would help address the issue of not running keepalived when the user intends not to use vip mode. #390
The text was updated successfully, but these errors were encountered:
The repo https://github.com/openelb/kube-keepalived-vip does not have issues enabled, so posting this here for now.
With the issues I ran into in #392 I considered going back to VIP mode given the HA provided by keepalived (when it doesn't keep crashing!), might be useful.
However in thinking about doing this I realized that there may be conflicts with
virtual_router_id
with multiple instances/clusters on the same network.This, among other customizations, I believe the template used for the keepalived.tmpl should be made available as a config-map in the cluster so the user can adjust as needed.
Also concerned about the flushing of IPVS tables when keepalived starts up, so access to the daemonset deployment template would also be useful, though I realize this could be edited after the cluster is deployed, with openelb managing that it feels there may be another mechanism needed to edit, or just leave deployment of that outside of openelb, which would help address the issue of not running keepalived when the user intends not to use vip mode. #390
The text was updated successfully, but these errors were encountered: