Hi all. I am having a little trouble with a particular scenario. I am quite new to JunOS so it's possible I'm not doing it right, hopefully somebody can set me straight!
The scenario is basically that I have an SRX1500 connected to a HP DL360 server running ESXi 6.5.0. I have four connections between them that I want to aggregate and trunk a number of VLANs across. This will need to be a static LAG as ESXi does not support LACP.
So, I have tried configuring it two ways, with mixed results. Firstly, I add the physical interfaces to the LAG:
ge-0/0/0 {
description Core
xf1nomacfelp01:nic1;
ether-options {
802.3ad ae0;
}
gigether-options {
802.3ad ae0;
}
Same for ge-0/0/1-3. Initially I just had ether-options, I added the second as part of my efforts to get it to work. Firstly, I tried create layer 3 interfaces on the LAG as follows:
ae0 {
vlan-tagging;
unit 100 {
vlan-id 100;
family inet {
address 10.10.10.1/28;
}
}
}
I then configured the management IP of the ESXi box to an IP in this subnet, configured vlan 100 and enabled all four NICs. I am unable to ping the box from the SRX, or anywhere else. However, if I disable three out of four interfaces on the SRX, I can ping it from the SRX and from my laptop. I have tried various combinations of enabled/disabled interfaces and found that sometimes I can ping with two interfaces enabled, or even three. All interfaces seem to work if enabled individually, just not together, implying there is some issue with the link aggregation.
The configuration options on ESXi are very limited so I can't see any reason to think the problem lies there.
The second thing I tried is to configure the ae0 interface as a layer two VLAN trunk, with the VLANs connected to layer 3 virtual interfaces. As follows:
ae0 {
native-vlan-id 100;
unit 0 {
family ethernet-switching {
interface-mode trunk;
vlan {
members device-management;
}
}
}
}
irb {
unit 100 {
family inet {
address 10.10.10.1/28;
}
}
vlans {
device-management {
vlan-id 100;
l3-interface irb.100;
}
With this configuration, I seem to have the same issues with the LAG, plus another possible issue with VLAN tagging. If I set the ESXi box to use vlan 100, it is not reachable. If I set it to untagged (plus add the native-vlan-id command to the SRX) it is reachable (again, subject to the issues with LAG.)
As I say, I'm pretty new to Juniper and I may well be doing it wrong. This particular scenario doesn't seem to have come up in the searches I have been doing on google, usually people connect an ESXi host to a switch rather than an SRX.
Cheers
Paul