Current Events > Juniper's access switches are freaking terrible, holy crap

(+) Yeah! (+)

Even though the proof of concept went terribly multiple times, we somehow decided to go with Juniper access switches any way. And sure enough, we're running to a bunch of bugs that Juniper has no explanation for.

On some of them, LLDP just stops working, which makes phones move to the wrong VLAN. We've also got alerts for missing VLANs on some servers, which depends on LLDP working, so we're getting false alerts.

Then we've got Juniper switches in virtual stacks and we've had multiple instances of members of the stack just deciding to leave the stack ring. We keep having to end out field techs to reboot the affected switches when this happens.

Then we've got an extremely widespread issue where 10 gig ports that are connected to Dell servers will just stop working. We've been hounding Juniper about this for a few months now and they've given us the necessary commands to go into the underlying Broadcom chip and reset each affected port one by one, but this is like a 35 - 40 hour job all by itself because of how many switches we have deployed.

And when we're escalating to Juniper TAC, it's like they don't have any answers to what is going wrong with any of this. We're definitely not the only company in the world experiencing this problem, but it's like we're the ones forcing them to figure the shit out.

I remember being excited to learn Juniper a year ago, but now I wish I had never touched this shit. It's awful.