Redis won't start via WebUI - Solved

[ThyOnlySandman]

Setup HA / CARP today on OPNsense 24.1.10_8.
LAN CARP VIP only and using single this DHCP WAN script to toggle master WAN INT
https://gist.github.com/spali/2da4f23e488219504b2ada12ac59a7dc
Outgoing NAT - LAN , Lan adresss --> LAN CARP VIP  (Source firewall traffic to use CARP LAN VIP downstream)
After a bit of struggle failover and sync is working.  (Make those OPT# match + administration - listen interfaces - All Recommended!)  IPSEC states don't though with this WAN toggle :(

Anyway - now noticing redis won't start.  Exact same error issue as discussed here
https://forum.opnsense.org/index.php?topic=38845.15

WARNING: The TCP backlog setting of 511 cannot be enforced because kern.ipc.somaxconn is set to the lower value of 128.

Warning: Could not create server TCP listening socket Real_Int_IP:6379: bind: Address already in use

Failed listening on port 6379 (tcp), aborting.

Just like other post.  I can CLI

redis-server --port 6379 --daemonize yes

It runs on same WebUI default port 6379 just fine.  Doesn't show running via WebUI.
And then can run Ntopng Enterprise via WebUI.

Other post was left without known cause...
Any ideas?  New HA setup shouldn't cause this right?

Edit:
Starting redis this way it is a fresh DB.
I haven't reset it via webui and HA sync shouldn't of reset it.  Only one way master to backup sync.  Hmm.

[ThyOnlySandman]

Post Bump.

I've ignored this Redis WebUI issue / bug since Aug. when I first post.

Today I upgraded Opnsense HA to 24.7.5
Went very smoothly.  Kudos to devs.

However still have this same Redis GUI bug with WebUI.  Won't start.  And strangely I'm not seeing any log errors now when attempting to start via WebUI like I had before on 24.1.

But same as 24.1 , Redis continues to work just fine if I use CLI -  redis-server --port 6379 --daemonize yes
However while running WebUI shows that it isn't.

Tried another redis reset + uninstall / reinstall + mv /var/db/redis /var/db/redis-OLD
No luck.

The exact same situation in previous post:
https://forum.opnsense.org/index.php?topic=38845.15

Any suggestions would be appreciated.

[ThyOnlySandman]

Interface binding issue as a result of HA setup - VIP / CARP IP.
Redis had been always set to LAN interface in past.

The reason CLI command -  redis-server --port 6379 --daemonize yes - had worked is because its implied loopback interface.  In Webui not selecting any redis interface also make it also use loopback and work.  I assumed like MGT binding it would mean it would listen on all interfaces 0.0.0.0 , not correct.  Just leave it with no interface selected.

With the new HA setup, either VIP itself and/or my LAN INT IP NAT config caused this issue.
I have all downstream traffic flowing through fortigate in transparent mode.  I have it using LAN VIP IP downstream, rather than LAN INT real IP of each firewall.

Whichever cause with VIP or my NAT, if Opnsense plugins are only thing to use redis - Don't select an interface and use loopback.