Bug #3255
closed
OpenVPN client or server on GWG does not failover
Added by Phillip Davis over 10 years ago.
Updated about 10 years ago.
Description
If I have a GWG "OpenVPNGWG" with WAN1=tier1, WAN2=tier2 then the initial OpenVPN client or server on "OpenVPNGWG" comes up correctly on WAN1. If WAN1 goes down, it does not failover to WAN2. Happens on systems with real NICs for WAN1 and WAN2, and systems with VLANs for WAN1 and WAN2.
Edit/Save of the OpenVPN instance will cause it to correctly evaluate the GWG state and select the best WAN.
- Status changed from New to Resolved
- Status changed from Resolved to New
Reopen it since pull request was not merged yet
Is the pull request, or some variation of the code in it, going to get merged some time soon?
- Status changed from New to Feedback
- % Done changed from 0 to 100
Tested with the 2.1 branch version of rc.openvpn
It successfully writes new servern.conf and clientn.conf files for the new highest-tier interface/gateway when the highest is unplugged. Writes them again when the highest-tier is reconnected. OpenVPN instances fail over and back. All good.
Now I need to get rid of "killed: out of swap space" and everything would actually restart nicely - but I can't blame the lack of memory in the Alix on this code!
- Status changed from Feedback to New
- Status changed from New to Feedback
didn't realize the merge was pulled, last update I noticed was Phil's 2 days ago, back to Feedback since it's merged.
My feedback above applies - using the version of rc.openvpn merged into the 2.1 branch. It is a good thing IMHO.
Ignore my comments about "killed: out of swap space" - just having a bad day with our main office Alix, not related to this code.
- Status changed from Feedback to Resolved
Also available in: Atom
PDF
Updated by 
Updated by 
Updated by 
Updated by Anonymous