We are not planing to change the backend to OpenVSwitch
- Feed Queries
- All Stories
- Search
- Feed Search
- Transactions
- Transaction Logs
All Stories
Today
Yesterday
Related task T7220
Hmmm, forgot to encode patch, sorry
We can close this as we we remove XDP driver from CLI (https://vyos.dev/T8202)
Sun, Jan 25
Yes, @GideonKuijt ; this is an intentional change to ansible-core from what I understand. They're trying to make the templating operations much more explicit, so some of the places where we were able to automatically have templating have been broken. Thanks for posting the solution here. I expect we'll see this in a few other places that have src: as well.
Sat, Jan 24
Minimal fix for the issues I described: https://github.com/vyos/vyos-1x/pull/4954
I think I've traced the issues. There are 3 different bugs:
Dev work is under https://github.com/vyos/vyos.vyos/pull/438
Fri, Jan 23
This is good - the problem seems more isolated now.
PR which contains an in-memory background operation manager that queues full configure operations as a single job:
Thu, Jan 22
Based on feedback from the PR, I will do further investigation, come up with a proper test case and comparison with how others (eg: OpenWRT) implement MSS adjustment.
FRR 10.5 removed support for applying route-maps to connected and kernel routes upon import from zebra. This is a breaking change, even though configs that used that feature are very unlikely to exist in the wild.
On further checking, I found that just starting the openhrp service brings the ipsec tunnel up, so it seems that opennhrp script takes the responsibilty of starting the tunnel in case of dmvpn and restart ipsec utility does not load the dmvpn profile.
@Viacheslav, the thread is closed and I can't find a way to message @hyyram on the forum. Can you or someone with enough rights on the forum, message him/her?
Reset command works but it will not be helpful if the tunnels are down from the start to re-initiate:
Wed, Jan 21
PR: https://github.com/vyos/vyos-1x/pull/4950 , although to be honest I am now perfectly happy with the implementation, there is likely a better way to implement it.
That's awesome - thanks
@zsdc After testing on our server this configuration works correctly (no dhcp address but still... no issues with MTU and static address):
yos@he-r650-01# ethtool -i eth3 driver: mlx5_core version: 6.6.117-vyos firmware-version: 16.35.4030 (DEL0000000016) expansion-rom-version: bus-info: 0000:31:00.1 supports-statistics: yes supports-test: yes supports-eeprom-access: no supports-register-dump: no supports-priv-flags: yes
@SrividyaA what about reset ?
I am happy to produce a pull request which adds the missing forward chain.