FortiClient VPN disconnect occasionally during remote session

[erosner]

Hello,

Very happy with the ForitClient VPN for the purpose of remote desktop to my office computer.

Lately, after updating the Client to version 7.0.1.0083 , I noticed that every time I leave my PC for few minutes (making me some coffee) when I return the VPN is disconnected. This is something new that didn't occur while using the previous version.

I attach a logfile with few occurrences of disconnection.

Basically what I see in the log file is:

10/1/2021 10:17:33 AM error sslvpn FortiSslvpn: 36320: error: poll_send_ssl ->SSL_get_error(): 5, try:1 10/1/2021 10:17:33 AM error sslvpn FortiSslvpn: 36320: error: poll_send_ssl -> [style="background-color: #ffff00;"]WSAGetLastError():2745[/style], try:1 10/1/2021 10:17:33 AM error sslvpn FortiSslvpn: 36320: error: poll_send_ssl ->data size: 51, try:1 10/1/2021 10:17:33 AM error sslvpn FortiSslvpn: 36320: [handle_driver_read_event]: error: poll_send 10/1/2021 10:17:35 AM info sslvpn FortiSslvpn: 26580: Ras: connection to fortissl terminated

Searched for this: WSAGetLastError():2745  and found this in FORTINET:

FortiClients can sometimes have connection issues with SSLVPN. This sometimes includes error messages as follows:

Error    VPN    FortiSslvpn: 2148: error: poll_send_ssl -> WSAGetLastError():2745, try:1 Error    VPN    FortiSslvpn: 10908: error: poll_recv_ssl -> WSAGetLastError():2746

These errors are Windows Socket event errors, expressed in hexadeximal format. WSAGetLastError():2745, is the equivalent of error code 10053.

And the error code 10053 in Windows socket error codes list:

WSAECONNABORTED10053Software caused connection abort.An established connection was aborted by the software in your host computer, possibly due to a data transmission time-out or protocol error.

So, it seems like something in my Host computer that disconnect the connection.

What should I look for? Any idea? 

Thanks for your help,

Eylon

[MarcAbaya1]

Has anyone figured-out what's causing this?

[DerekWSmall]

Has anyone gotten anywhere on this?  I have a customer with Verzion Buisness as their primary ISP.  I have trouble VPN'ing to several other customers when I am onsite and using their Verizon Internet.  They have Fortigate Firewalls as well.  If I do packet captures, I the TCP three-way hand-shake, then a FIN-ACK and an ACK, then nothing.  The Forticlient complains about a TLS version mis-match, but the packet capture shows that we don't even get to the TLS negotiation.  I've tried several versions of FortiClient, the latest being 7.0.6.

I'm trying to VPN in from a VMware VM running on my laptop.  If I switch my laptop to a tethered connection on my Verizon cell phone, I have no problems.  But if I try to bring up the VPN tunnel while my laptop is on the customer network, it always fails at about 40%.  

[PierrePaulDuval]

Thought I would chime in.  I have some insights regarding this issue.  I work for Retraite Quebec and we recently switched our VPN client, Cisco AnyConnect, to FortiNet.  Unfortunately, we started with version 7.0.7 which was an almost horrible experience with the VPN (all the other parts are fine, thank god):
- VPN : Very frequent disconnecting for almost all users;

- MFA stopped working (getting notifications or insert pin);

- Prompting users for admin priv. when starting Forticlient;

- For those who have "ConnectBeforeLogon", the shield (credential provider) would show at Windows login, so no connectbeforelogon.  Users logs in with Windows as cred. prov. then connects to VPN;

- Getting toast notified 150 TIMES a day telling me there is a problem with the FortiClient addon in Edge and I need to reinstall it.   That was a PAIN.  I simply turned it off via GPO.

Don't get me wrong, it's a cool product, speed is very nice, big upgrade from Cisco and ConnectBeforeLogon fixed at least 5 or 6 things because the user has now access to the network.  

BUT 

The more I read about this issue (frequent vpn drop) on forums, the more I realize that this didn't start last month or last year.  It's been years.  We tried a LOT of things.  But ultimately, we decided to install the "latest but not really latest version : 7.0.8".   I say this because 7.2 came out before 7.0.8.  So this tells me that issues in 7.0.7 were significantly important to patch it AFTER 7.2 came out. 

Here's one great advice:  Update your Wifi device drivers and disable the MIMO power save mode
https://www.intel.ca/content/www/ca/en/support/articles/000005585/wireless/legacy-intel-wireless-pro...

Why disable?  Because all antennas on your wifi card will be powered and ready when your PC or laptop is in all the other state than normal :).  I guessed this could help and it did.

After upgrading to 7.0.8 and updated all wifi devices driver (plus disabled powersavemode), the VPN drops went wayyyy down but we still do have some occasionally. 

If there is one thing Cisco was very good at : Incredible stable connection.  I miss that.  Am I the only one that doesn't think it is normal to see your VPN break when you change your WiFi connection to another one (ex: 2.4 to 5Ghz) ?

Thanks!

[ivanov_ip]

Can confirm! Your solution does work, and fixes the issue!

For everyone else: just change Auto SMPS mode to No SMPS in MIMO power save mode setting inside Advanced device configuration properties (you can reach this window through Device Manager -> Network adapters or Settings -> Networks -> Wi-Fi -> Hardware properties). That's all!

[jasnjay]

has anyone have a solution on this? we are also encountering this issue

[Flipper-One]

Its 2024 April with FortiClient v7.2.3.0929

Error: WSAEnumNetworkEvents FD_CLOSE (10053)

MS Wifi Direct Virtual Adapter #2 on the laptop was turned off/disable as a workaround seems work for a while but it he issues returned agany.

Any clue what causes it?