FortiSwitch disconnected but operational

[JPMfg]

Hi,

There have been a few similar posts, but none with the same constellation (OS level and topology):

We have a single FortiGate 100D running FortiOS 5.6.3 managing a stack of two FortiSwitch 124E with S124EN-v3.6.3-build4269.

Connection is:

FortiGate FortiLink LAG using Ports 12 and 13 connecting to Ports 23 and 24 of switch #1 (copper, no split-interface).

ISL (fiber optic) between Switch #1 and Switch #2 on ports 25 and 26 (25 on 25 and 26 on 26). (no direct link between Switch #2 and the FortiGate possible).

Switch #1 shows as "disconnected" and cannot be configured anymore. Switch #2 is functioning fine.

We are going to schedule a maintenance window and reboot the switches, but i would like to know how this could have happened and how to prevent it in the future.

Diagnostic: I can ping and telnet to both switches using the automatic FortiLink addresses (169.254.1.x), the disconnect happened more than 7 days ago (it's a new environment, no monitoring yet) so the event logs doesnt show anything anymore.

Diag status on switch #1 just confirms that there is no connection. Configuration is otherwise ok.

Since the Switch #2 is working correctly even though it is connected through Switch #1, i do not want to do any disruptive debugging on the "disconnected" switch. Is there any way to re-start the connection that does not disrupt connectivity for the connected devices?

[gummi]

Hi 

I have in a few cases killed the switch manager service on the switch which does not want to connect and after that they have connected 

can be done with 

diag sys top    -- and get the pid of the cu_swtpd service which in this case was   ‚cu_swtpd      923‘Then kill the service with – ‚diag sys kill 9 923‘ regardsGudmundur

[JPMfg]

gummi@sensa.is wrote:

can be done with 

diag sys top    -- and get the pid of the cu_swtpd service which in this case was   ‚cu_swtpd      923‘Then kill the service with – ‚diag sys kill 9 923‘ regardsGudmundur

Does that affect traffic?

[chilimac02]

Did you find a solution to this issue? I'm having a similar issue myself. Even a link to another post would be helpful :)

[bommi · Accepted solution]

Once my switch had an wrong time after an software upgrade (01-01-1970) in that case the management connection did no come back.

I checked the arp cache of my fortilink port to find the actual ip of my switch:

# diagnose ip arp list | grep 169.254

index=8 ifname=lan4 169.254.1.2 XX:XX:XX:XX:XX:XX state=00000004 use=708 confirm=3118 update=64 ref=3 Then I logged in using ssh and set the correct time:

# execute ssh admin@169.254.1.2 # execute date <yyyy-mm-dd>

Some seconds later the connectivity has been established.

Regards

bommi

[tanr]

I created an additional post about the 124E FortiSwitch that referenced these problems and wanted to check if (and how) the problems have been resolved, or if they haven't been resolved.  Please let me know.

Link to the other thread:

https://forum.fortinet.com/tm.aspx?m=163337 

[JPMfg]

chilimac02 wrote:

Did you find a solution to this issue? I'm having a similar issue myself. Even a link to another post would be helpful :)

We rebooted both fortiswitches and the fortigate during a maintenance window and also upgraded the firmware.

This removed the issue (after reboot both switches are now active).

[Kwest]

We rebooted both fortiswitches and the fortigate during a maintenance window and also upgraded the firmware.

This removed the issue (after reboot both switches are now active).

My issue persists! Even so with all upgraded and rebooted... :(