# Random disconnects and ignoring auto reconnect setting

**URL:** <https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555>\
**Category:** Pritunl VPN\
**Tags:** pritunl, pritunl-client\
**Created:** [October 29, 2024, 2:16am UTC](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555 "2024-10-29T02:16:18Z")\
**Posts on this page:** 4\
**Page:** 2

<div class="post-metadata">

**Author:** ![jaredc](https://forum.pritunl.com/letter_avatar_proxy/v4/letter/j/dec6dc/32.png) [@jaredc](https://forum.pritunl.com/u/jaredc)\
**Post date:** [December 10, 2024, 6:03pm UTC](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555/21 "2024-12-10T18:03:17Z")

</div>

I think I have tracked down my issue

my connection is being killed with this log on the server

[eve-9199] 2024-12-10 12:51:17 wg-client-kill “auth\_update\_err”

This indicates the auth failed, which makes sense as we use radius with 2fa code added with the password.

My query is the client has no indication that the connection was interrupted in the logs , so why is the reauth happening? Wireguard by nature is connectionless , is this related to the WG ping and timeout setting in the server?

Auto reconnect is disabled, yet an auth was sent to the radius server.  
This affects our 2fa keys which have a max fail counter

is there a doc page that explains the pritunl wireguard connection lifecycle

Edit: what confuses me more is I can simulate a “connection break” by switching to my 2nd isp , but this does not trigger the above reauth

---

<div class="post-metadata">

**Author:** ![zach](https://forum.pritunl.com/user_avatar/forum.pritunl.com/zach/32/1105_2.png) [@zach](https://forum.pritunl.com/u/zach)\
**Post date:** [December 10, 2024, 9:45pm UTC](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555/22 "2024-12-10T21:45:29Z")

</div>

There is connection management in the Pritunl client and server for WireGuard connections. This is documented in the [**WireGuard documentation section**](https://docs.pritunl.com/docs/wireguard).

For both OpenVPN and WireGuard connections an authentication verification is done every hour. This would disconnect existing connections in the event that the single sign-on user no longer had valid status with the single sign-on provider. This can be disabled by running `sudo pritunl set app.sso_connection_check false`.

---

<div class="post-metadata">

**Author:** ![jaredc](https://forum.pritunl.com/letter_avatar_proxy/v4/letter/j/dec6dc/32.png) [@jaredc](https://forum.pritunl.com/u/jaredc)\
**Post date:** [December 11, 2024, 1:57pm UTC](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555/23 "2024-12-11T13:57:18Z")

</div>

This setting seems to have fixed it!

Thanks for the info

Does radius authentication get affected by:  
app.sso\_cache  
app.sso\_client\_cache  
app.sso\_cache\_timeout  
app.sso\_client\_cache\_timeout

I would need to adjust these to align with our security policies etc

---

<div class="post-metadata">

**Author:** ![zach](https://forum.pritunl.com/user_avatar/forum.pritunl.com/zach/32/1105_2.png) [@zach](https://forum.pritunl.com/u/zach)\
**Post date:** [December 11, 2024, 11:36pm UTC](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555/24 "2024-12-11T23:36:32Z")

</div>

Radius will be checked when connecting even when cache is enabled.

[Previous page](https://forum.pritunl.com/t/random-disconnects-and-ignoring-auto-reconnect-setting/2555.md?page=1)