Cluster Interface Is DOWN

[JohnnyWang]

Dear Guy,

Currently we are facing the issue related to Cluster Interface is DOWN on Standby Cluster. We have bond10 on the Gateway, and interface bond10.48 is DOWN on the Standby Gateway. Another interface in this bond is ok, but only bond10.48 is DOWN, and Another Gateway is ok. If, we failover HA to this server, it will show ACTIVE (F) status.

Please suggest for me if everyone facing issue the same.

[Chris_Atkinson]

Have you checked to confirm all is well at Layer-2 / VLAN level on that trunk?

Which Version & JHF is this cluster?

[JohnnyWang]

Hi @Chris_Atkinson ,

Yes this is l2 interface in trunk port with switch. I have check on switch, they have added this vlan on the trunk port. Another firewall in cluster is UP but I don't know why in this gateway is DOWN.
Gateway version r82 hf 126

[Chris_Atkinson]

So everything looks as expected when you do a packet capture on that gateways interface?

Please share additional cphaprob & bond cmd outputs... e.g.

show bonding groups

cphaprob show_bond

cphaprob show_bond bond10

[JohnnyWang]

Hi Bro, 

Below is Info request

Bonding Interface: 10
Bond Configuration
xmit-hash-policy layer2
down-delay 200
primary Not configured
lacp-rate slow
mode 8023AD
up-delay 200
mii-interval 100
min-links 0
Bond Interfaces
eth2-01
eth2-02
eth3-01
eth3-02
Bonding Interface: 20
Bond Configuration
xmit-hash-policy layer2
down-delay 200
primary Not configured
lacp-rate slow
mode 8023AD
up-delay 200
mii-interval 100
min-links 0
Bond Interfaces
eth2-03
eth3-03
Bonding Interface: 40
Bond Configuration
xmit-hash-policy layer2
down-delay 200
primary Not configured
lacp-rate slow
mode 8023AD
up-delay 200
mii-interval 100
min-links 0
Bond Interfaces
eth1-07
eth1-08
cphaprob show_bond

|Slaves |Slaves |Slaves
Bond name |Mode |State |configured |link up |required
-----------+-------------------+------+-----------+--------+--------
bond40 | Load Sharing | UP | 2 | 2 | 1
bond10.30 | Load Sharing | UP | 4 | 4 | 3
bond10.615 | Load Sharing | UP | 4 | 4 | 3
bond20.220 | Load Sharing | UP | 2 | 2 | 1
bond20.901 | Load Sharing | UP | 2 | 2 | 1

Legend:
-------
UP! - Bond interface state is UP, yet attention is required
Slaves configured - number of slave interfaces configured on the bond
Slaves link up - number of operational slaves
Slaves required - minimal number of operational slaves required for bond to be UP

[Chris_Atkinson]

Unless you have implemented parameters like those in sk92826 only the highest and lowest VLANs are actively monitored.

Do any of these outputs differ across the cluster members, any clues in /var/log/messages ?

Are you already working on this issue with TAC?

[emmap]

What does 'show interface bond10.48' say about it?

[ZaferGr]

Hi,

Can you share the output of this command from all gateways?

more /proc/net/bonding/bond*  | grep MII

 Also, are you sure the other end is configured correctly?

[JohnnyWang]

I am sure it is correct; it has been working since 2024.

[ZaferGr]

Hi,

In that case, draw up a roadmap,

First, you need to check the VLAN configurations on the switch connected to the gateway that’s down.

If there’s no issue there, I recommend reconfiguring the bond on the problematic gateway. You can even do this on different interfaces.