Local connections dropped at seemingly random intervals

[greffter]

I am experiencing a perplexing, and troubling problem where at seemingly random intervals all devices on my local network lose connection to the router.  I have no idea if WAN connectivity is dropped as well since, during these moments, my internal IPs are dropped.  This started happening ~2 days ago which roughly corresponds to updating to OPNsense 21.1.1.

Topology

Fios -> OPNsense 21.1.1-amd64 running on a 6 port Protectli -> Cisco managed switch

Connections in use

- Lagg0 (three nics using LACP in trunk mode to the switch) -> 10.0.2.0/24
- VLAN10 (HOME) -> 10.0.0.0/24
- VLAN 20 (Guest) -> 10.0.20.0/24
- VLAN 30 (Servers) -> 10.0.30.0/24

When the connection problems occur, I receive an IP from the opnsense in a range I'm not using at all.  192.168.97.15 with a netmask of 255.255.0.0

Unfortunately, I'm really not certain where to start even troubleshooting this issue.

[greffter]

Providing a little more information

There were three changes made the day that I started seeing these errors:

1. upgrade to OPNsense 21.1.1-amd64
2. Move /var and /tmp to a RAM disk via the GUI
3. Blocking all IPv6 on the entire network

Hopefully that information may help someone point me in a direction to start to troubleshoot this

[greffter]

Following up here in case anyone else seems to encounter this problem.  It seems to be correlated to the usage of RAM disks for /tmp or /var.  Disabling those settings in Administration > Misc cleared up the problem for me.

Does anyone know why these settings may be causing periodic outages of OPNsense connectivity?

[malakez]

After upgrading to 21.1 I lose LAN connection to the router after 10-15 minutes so I had to downgrade to 20.1. My network

VDSL-router in bridge mode -> shuttle ds77u (opnsense) -> uqibuiti switch -> PC

But when I connected

VDSL-router in bridge mode -> shuttle ds77u (opnsense) -> PC

it still dropped connection to Shuttle after 10-15min to shuttle. First dropped WAN connection and few minutes after that LAN-connection. So I cannot update to 21.1 even if i wanted to :(

[speddero]

Hi Folks.... I started using Opnsense about 6 months ago when I made the open source leap given all the great things I read.. been absolutely loving it ever since - awesome work OPNsense!  However I've been experiencing what appears to be a similar issue as described in this post.   This has only started right after I upgraded to 21.1.1 one week ago.  In my case, all traffic gets dropped on any client in my network (can't tell if it's WAN or LAN).  It is regular and predictable.... every 15 minutes EXACTLY.  I'm typically on video calls all day, and used the timer on my phone every time I 'freeze'... coupled with a running ping to google.com which also fails every 15 min to validate the timing.  I noticed it on other devices as well at the same time (including my netflix on smart tv).  I've checked all the logs that know of via the GUI - nothing pops out as relevant or obvious and nothing with that 15 minute regularity.  Nothing running in cron either.  I'm not BSD fluent and don't know what else to look for.  It's a pain for work conf calls, coupled with kids in virtual school etc.  I'm running a PCEngines APU2D4 hardware.  Not sure if it's a kernel thing, NIC thing.... just not smart enough yet to know what to look for.  Any help would be appreciated.. even if it's instructions to roll back versions (tried rolling back using GUI, but that appears to only roll-back config settings).  Many thanks in advance.

[speddero]

Sorry... forgot to also mention the dropped connectivity only lasts for ~10-20 seconds every time... again repeatably and reliably.

[speddero]

One more thing... I did check my settings per the earlier part of the thread regarding /tmp and /var ramdisks... both of these are unchecked in my configuration. 

My ISP is a WISP.... outside radio router on tower (bridge mode) -> OPNsense (PCE APU2D4) -> devices

I've also reached out to my WISP as well.