Link-Monitor broken in 5.2.3 ?

[TC_Hessen]

Hi,

I found an extremely strange behavior in the FortiOS 5.2.3 on our FGT 110C.

FG-GW # diagnose sys link-monitor status

Link Monitor: gwdetect-upg-1 Status: alive Create time: Sun Mar 29 01:17:53 2015
Source interface: wan1 (3)
Interval: 5, Timeout 5
Fail times: 0/5
Send times: 0
[...]
    protocol: ping, state: alive
[...]

Link Monitor: gwdetect-upg-2 Status: alive Create time: Sun Mar 29 01:17:53 2015
Source interface: wan2 (10)
Interval: 5, Timeout 5
Fail times: 0/5
Send times: 0
[...]
    protocol: ping, state: alive
[...]

seems to be correct, but

FG-GW # diagnose sys link-monitor interface wan1
Interface(wan1): state(down, since Sun Mar 29 01:18:21 2015
), latency(8.96), jitters(18.33), session count(639), bandwidth(55977)

FG-GW # diagnose sys link-monitor interface wan2
Interface(wan2): state(down, since Sun Mar 29 01:18:22 2015
), latency(21.22), jitters(8.04), session count(81), bandwidth(30896)

show's down links. The same output are shown in the gui on System/Monitor/Link-Monitor. Is this my failure or should I report an bug?

[Dave_Hall]

We don't have any fgts on 5.2, but have been thinking about this problem for the last 3 days.  I am just wondering if one command just confirms the service/monitor is up/active while the other command confirms the link itself is either up or down. 

Since ping is being using, I can imagine some funky results if both WAN1/WAN2 connections are up, but not the ping servers.

Otherwise, I think it is a bug.

[msigl]

Hi,

i can confirm this also. I have a client with 5.2.3 with 80D and another with 30D. Both show on GUI link monitor the status down. Can confirm also that from CLI from the 1st command i get the links alive, and with the 2nd command the links are down.

Probably a bug. Found this in release notes > Bug ID: 268019 The VWL and link-monitor status may not be correctly indicated in the GUI.

Kind Regards,

[luca_vasini]

same problem...

what will come out the fix?

Thanks

Luca

[Christopher_McMullan]

It's still being tested. Best to open a support ticket with Fortinet TAC to have your situation associated with the bug, so you can be made aware of firm patch release dates as they become available.

[JayWinksInLine · Accepted solution]

It gets better. The behavior I'm seeing is that the display under System -> Monitor -> Link Monitor in the UI will always be precisely opposite the output of di sys link-mon stat for the corresponding monitor.  The CLI output is correct, the UI is always reverse and as such wrong.

   

[aseques]

Any news on this? I have a unit on 5.2.3 with the same issues and would like to know if it's fixed on newer releases?

[aseques]

Answering myself, it seems that it should be fixed now, I see two mentions of this in the release notes for 5.2.4.

[ul]

268019 The VWL and link-monitor status is not correctly indicated in the GUI.

278350 Link monitor event log message is misleading.[/ul]

[Luiz_Alberto_Camilo]

Well, I'm testing exactly the same scenario and getting exactly the same behaviour. 

I'm using the latest 5.2.6 build and when the link monitor brings down an interface, even if the link monitor comes back up, the interface remains down. 

As a workaround, if you run a "exec router restart" the interface comes up again after a few seconds. 

Problem not solved. 

[TC_Hessen]

No, we are using a 60D and a 110C cluster with 5.2.6 on it. Since the firmware 5.2.4 fixed this bug, it does not come up again.