webgui: long lasting calls and locking the request pipeline. work-around for https://github.com/opnsense/core/issues/7651
The number of php processes we offer to lighttpd is quite low (2 x 3), which results in a global lock when 2 dashboards are open at the same time.
Although we should increase the limit, we also should figure out a way to limit the number of long running processes as well.
(cherry picked from commit 724f8494d64074e0384b9f593e699daf2d84d604)
webgui: long lasting calls and locking the request pipeline. work-around for https://github.com/opnsense/core/issues/7651
The number of php processes we offer to lighttpd is quite low (2 x 3), which results in a global lock when 2 dashboards are open at the same time.
Although we should increase the limit, we also should figure out a way to limit the number of long running processes as well.
VPN: OpenVPN: Servers [legacy] - disable DCO, only supported for new instances (and prevents server startup for tun devices).
(cherry picked from commit 89135cdc7bf23ad57ca9fcd725b2b13989a3aea1)
Interfaces: Virtual IPs: Settings - polish diff calculation a bit when using the new peer settings (avoid re-apply when unchanged)
(cherry picked from commit 95812e27864f44c681ae883278e846d6ea25c5a2)
system: fix disk widget glitching during install
"B" unit was not properly handled during multiplication. ;)
(cherry picked from commit 94a9231a8f89b1e24e2e3100cc0856b170fc9e40)
Firewall: NAT: One-to-One - fix migration issue, in the old setup external without a subnet wsa also a valid option...
(cherry picked from commit cca9149e05b99478e857732d4327cd6cc34b9d0c)
Firewall: NAT: One-to-One - fix migration issue, in the old setup external without a subnet wsa also a valid option...
(cherry picked from commit cca9149e05b99478e857732d4327cd6cc34b9d0c)
VPN: OpenVPN: Client Export - dev_type should be constraint to either tun or tap, DCO uses ovpn, which equals tun for the client
(cherry picked from commit 228ccc0722a5b04ed5ef6927e0f90b2895065a96)
interfaces: refuse to set an empty MAC address
This goes for the LAGG claming their original hardware address
is zeroed which is true but not really helpful at all, but also
for when this MAC address was requested by the user in the settings
which likely makes it stop working altogether anyway.