2.5.2 Wan disconnects

[jeffsmith82]

I upgraded to 2.5.2 the day after release on a pair of servers in a HA setup and for some reason the WAN interface on the primary keeps going down after about 36 hours. It's igb5 in the log below.

Jul 10 10:26:31 fire01.internal.XXX.com filterlog[13998]: 4,,,1000000103,igb5,match,block,in,4,0x0,,247,38690,0,none,6,tcp,40,193.27.228.63,83.126.54.36,44725,9775,0,S,575113001,,1024,,
Jul 10 10:26:33 fire01.internal.XXX.com charon[14898]: 11[KNL] 172.16.98.1 disappeared from igb1
Jul 10 10:26:33 fire01.internal.XXX.com kernel: carp: 4@igb5: MASTER -> INIT (hardware interface down)
Jul 10 10:26:33 fire01.internal.XXX.com kernel: carp: demoted by 240 to 240 (interface down)
Jul 10 10:26:33 fire01.internal.XXX.com kernel: igb5: link state changed to DOWN

If I reboot the server it comes back up and becomes the master again without issue. Any idea what could be causing this ?

[Gertjan]

@jeffsmith82 said in 2.5.2 Wan disconnects:

Jul 10 10:26:33 fire01.internal.XXX.com kernel: igb5: link state changed to DOWN

Most probably, general cases : the device connected to NIC igb5 brought the interface down.
Or : the cable is bad.
Or one if NIC's involved is bad.

This is a HA set up.
Maybe the interface is brought down for some 'HA' reason, but I can't tell, never used HA before.

[jeffsmith82]

@gertjan said in 2.5.2 Wan disconnects:

ably, general cases : the device connected to NIC igb5 brought the interface down.
Or : the cable is bad.
Or one if NIC's involved is bad.
This is a HA set up.
Maybe the interface is brought d

ITs only started happening after 2.5.2 upgrade so its possible bad nic or cable but would have thought restart wouldn't have fixed it.

[jimp]

It's possible the hardware problem started from the reboot during upgrade and not from anything in 2.5.2.

You'd be surprised at how many hardware issues only start or get noticed after a reboot, and then seem to coincide with an upgrade as a result.

But there isn't a lot to go on to say with any certainty what the cause might be.

Are those logs in reverse order with the newest on top? It appears that way.

Normally that kind of log message about a NIC only happens from a hardware event (e.g. cable is unplugged).