Catalyst Crash - 17.9.x after onboard

Lurick
Getting noticed

Catalyst Crash - 17.9.x after onboard

Just wondering if anyone else has onboarded a 9200/9300 running 17.9.1 or 17.9.2 successfully without a crash? I tried earlier on a lab device and it seemed to work for about an hour before it crashed and the switch reloaded. I suspect it's got something to do with repeated SSH access/logins to the switch from the dashboard but not 100% sure and want to see if others have had similar behavior on these 17.9 releases or even 17.6. I know there is a bug with Tacacs and intermittent connectivity but I don't have Tacacs configured on this device so I'm thinking that should rule that out.

8 REPLIES 8
cmr
Kind of a big deal
Kind of a big deal

We have a stacked pair of 9300s running 17.6.3 and they have been fine, other than the port channels getting reset in the 'glitch'.  The other oddity is that ports don't always stay green on the dashboard, particularly those on the 8-port 10G module that are members of a port channel...

SVTWeb
Here to help

As part of our Meraki-Catalyst PoC we have tried everything from 17.3.1 to 17.9.2. 

 

Thus far our findings on the 17.9x have been consitant with severe instability on the access layer swtiching. We have not posted internally that 17.9x is not to be used in a production environment. 

 

17.6.3 & 4 has thus far been stable in our environment.  We have however hit a snag in the last 24 hours that as we dig deeper as all of our PoC switches have lost coms with the Meraki cloud. 

Lurick
Getting noticed

That sounds good, glad to hear it's not something that's shown up on other trains/versions either. What's really odd is I have a 9300 on 17.9.2 which has worked fine for a couple weeks now without issue. I'm going to dig into the 9200 in the lab once I get a chance as there might be something there that's causing this, which would be nice to figure out.

Edit:

I can confirm the 9300 has also lost Meraki cloud comms in my lab as well.

the 9300's acting as a core have not shown issues with 17.9.x in our world either.. but we make it a point to ensure that all hardware in aligned on the same tested and validated version. That was the reasoning behind blacklisting the 17.9.x versions at this time. Better safe than sorry. 

cmr
Kind of a big deal
Kind of a big deal

@SVTWeb did your switches come back?  Ours were only offline (from a Meraki point of view) for 10-15 minutes

SVTWeb
Here to help

Still showing offline...  I can see telemetry going out over our firewalls without any issue... just not showing up in the dahsboard. 

SVTWeb
Here to help

Just tried to onboard some other sites and they fail at the "Checking Device Eligibility" task...  Appears something is no happy on Meraki's end... 

SVTWeb
Here to help

IT'S ALIVE!!!!!!!!  Just came back to life! 

Get notified when there are additional replies to this discussion.