MX100/MS120 Auto Negotiation Issues

[Mloraditch]

I've experienced this numerous times, but when I do, it's almost  never in a state where I have time to deal with support.

We use MS120-8s as DMZ switches. Sometimes when they or the MXs do a reboot/firmware upgrade the WAN Ports do not come back online. If somehow your MS120 is still online you can change the ports to 100/full, which "kicks" them online and then once online change back to auto and you are good to go.

Interestingly enough: https://bst.cloudapps.cisco.com/bugsearch/bug/CSCvu65369/?rfs=iqvred this bug was recently filed on the main Cisco side of the house.

Has anyone else seen this? Did you come up with a solution? It's only MX100s I've seen this in. I can't hardcode to 100/full as my bandwidths are larger than that and while the MS120 allows a hardcode to 1gb, the MX doesn't.

[cmr]

I always use dumb L2 unmanaged switches on the WAN ports and we don't have the issue you described when we reboot our MX100s.  We use the MX15 firmware train and keep it pretty up to date so they are rebooted about once a month or so.  I can confirm that they always come back at 1Gb FD.

[Mloraditch]

At a good number of sites we have third parties that also need connections to the DMZ so we need to be able to monitor them. Hence the MS120s.

It appears dumb switches are smart and smart switches are dumb in this scenario!!