Juniper vQFX issues with LACP/MC-LAG

Before posting something, READ the changelog, WATCH the videos, howto and provide following:
Your install is: Bare metal, ESXi, what CPU model, RAM, HD, what EVE version you have, output of the uname -a and any other info that might help us faster.

Moderator: mike

Post Reply
rsxhawk
Posts: 13
Joined: Tue Sep 11, 2018 4:19 am

Juniper vQFX issues with LACP/MC-LAG

Post by rsxhawk » Wed Sep 12, 2018 2:48 pm

I recently performed a bare-metal installation of EVE-NG on an HP DL360p G8 server and thus far everything has been performing great. I was able to successfully provision two vQFX machines and I was able to get LACP working between them with no issues. But several hours later, when I tried testing connectivity again, I noticed that LACP had broken even though "show chassis fpc" still showed the PFE online and testing. Sometimes rebooting the RE and PFE fixes this, but I am curious if there was something else I can do to maintain the connectivity? As of this morning, LACP is working again and I have started a continuous ping between vQFX1 and vQFX2 over the ICCP link (ae0).

I am using the older 15.1X53-D60 RE image and the 20160609-2 PFE image.

Server = HP DL360p G8 with 2x Xeon 2640 CPU's, 64GB of ram, 1x 256GB SSD and 1x 1TB SSD
Bare-Metal EVE-NG install on Ubuntu 16.04.5 LTS

Additional Information:
I am trying to test MC-LAG with two Ubuntu VM's connected via bonded (LACP) interfaces but I'm not sure what the back end requirements would be to accomplish this. I have set up bond0 interfaces on the linux hosts and created a LAG on the vQFX's, but I don't get any connectivity. I noticed that the only interfaces in the EVE gui I can select on the Ubuntu VM's are the regular ens3/ens4 interfaces, not the bond0 interface I created inside the Ubuntu VM. Is there a best-practice way of accomplishing this? Do I need to use a different version of linux?

Post Reply