Meraki

Nash · ‎Jan 23 2020

Even if ECMS2 isn't at CLUS, there will be other events and trainings related to general network goodness and Meraki in specific. I went in 2018 and there was quite a bit of Meraki or Meraki-related stuff, especially if you were interested in network automation.

Nash · ‎Jan 23 2020

I have attended, but it is very difficult to get into. My classmates were primarily Cisco sales engineers, large Meraki partners, and mostly otherwise people from companies with millions of USD in Meraki gear. Some of whom surprised me... I had no idea those companies used Meraki! There were also only 10-12 students. When they say seating is limited, they mean it. We had two trainers in a room for 3 days of training, plus however much time setup/breakdown takes them. I hear you that it's frustrating, But it's largely a matter of resources. Meraki's training division is only so large, and ECMS2 is just one of the things that they do.

Nash · ‎Jan 16 2020

@NFL0NR wrote: when in doubt... reboot. We've had issues today with 3 different devices that rebooting them resolved. As a matter of best practice, I wouldn't do this as an initial step. If you reboot, you lose the local logs on the device. Sometimes support needs to review those when you call in about a problem.

Nash · ‎Jan 16 2020

Tracert is an entirely different protocol from HTTPS. We're talking ICMP vs TCP. What happens when you run a pcap on your Internet interface while attempting to go to Google? And on your LAN? That will show you what's happening with the TCP connection on 443.

Nash · ‎Jan 16 2020

What kind of connection are you talking about? From a Meraki device to the Meraki cloud? From MX to a site-to-site VPN? From internal device? A pcap will tell you a lot about what's actively going on.

Nash · ‎Jan 16 2020

I recall the issue in 2017. But there's no reason to assume the same problem is happening. What happens when you run a pcap while attempting to get to Google.com? What's the traffic show? I'd probably dump it into Wireshark to review, btw, since you'll need to look at both the LAN and Internet interfaces and that's a good bit of traffic.

Nash · ‎Jan 15 2020

New people! Congrats, everyone!

Nash · ‎Jan 14 2020

If you can't change the routing protocol in use to OSPF or convert to static routes, you're not going to be able to use the MS250. EIGRP is Cisco's special sauce. If EIGRP is required, I'd see if you can get your money back.

Nash · ‎Jan 14 2020

Oy, that sounds frustrating. Too bad you can't lock them in a cage. Perhaps one that provides an incentive to not touch it. (Such as some nice punky spikes, or it makes a squeaky chalkboard sound no matter what touches it.)

Nash · ‎Jan 13 2020

Then it's repeating information across calls, over and over. I tend to revert to database thinking on this. The organization ID, network ID, and serial number function like primary keys and foreign keys. Using those keys, you can access specific information that all tie back to those keys. If you need the name of the appliance, not just the serial number, pull a list of devices for that org or that network. Then use the device serial number as a key to connect name to uplink status.

Nash · ‎Jan 7 2020

The acronyms are so similar, it's designed for confusion.

Nash · ‎Jan 7 2020

Certified Meraki Networking Associate (CMNA) isn't deprecated. It's a Meraki partner-specific training, with trainings available in January or February depending on region. None close to me this round, so I'm glad my official mentee already went through it! If you're a partner, run don't walk to go look at https://www.merakipartners.com/ to see if there's a session in your region. I believe they were just updated this morning. Certified Meraki Network Operator (CMNO) is no longer a separate cred, tho, since ECMS1 replaced it. My partner took ECMS1, since his office runs a full Meraki stack now, and enjoyed it a great deal. I've done ECMS2, and enjoyed it more... even if I was allergic to my hotel room...

Nash · ‎Jan 4 2020

Congrats, everyone! I was actually looking for (and could not find!) that thread that @TimBecker started, recently! Now I've got an easy place to locate it, hah.

Nash · ‎Dec 31 2019

If you run into a specific issue, post here and we'll see what we can do for you. None of us started off knowing how to make it work.

Nash · ‎Dec 21 2019

Yesssss the water bottles are so good! I’m glad to see I’m not the only person in loooove with theirs. Hard agree that general network knowledge is essential to really using Meraki kit effectively. Sure, it’s easy to deploy a bunch of switches, but you’re going to have a bad time if you don’t understand VLANs and STP. Congrats on the focus, @Priesty!

Nash · ‎Dec 18 2019

Make sure you've got a software firewall enabled on your laptop before you run this test!

Nash · ‎Dec 17 2019

Oh, that's interesting in a bad way. What happens if you run pcaps off your primary WAN interface and off LAN? I'd save them as a proper pcap and review in Wireshark. What traffic do you see?

Nash · ‎Dec 17 2019

I've also requested to have a network moved to a different registry, when the MX kept losing connectivity to the registry.

Nash · ‎Dec 17 2019

@BrechtSchamp wrote: I knew I was going to get hit with counter examples. Curse you @PhilipDAth 🧛🏻‍ Why don't you fetch the orgs and networks you need directly by using their ID though? The point I'm trying to make is that you have to be careful with them as it's easy to make mistakes. Your script will be running in an unexpected way and you might unwillingly be skipping important actions. And your script can be harder to understand as the flow is kind of interrupted. This is where logically-designed pseudo code comes in for me. I write it to help my thinking, then adjust it into comments on my code. No code is self documenting. I will fite anyone who argues. 🤣 Yes, I absolutely sometimes don't comment my code as much as I should.

Nash · ‎Dec 16 2019

@BrianP wrote: Fair enough. For people who have issues hitting the Rate limit ( due to other active monitoring API calls ), it would be significantly faster to do the network list call and the full org device list call, and correlate the device to the correct network in the code, instead of the individual calls per network. Either way, you solved Raphael's issue! Definitely agree. 🙂 I can do things sloppily sometimes because Auvik and I are the only things making calls, and Auvik rate limits itself. There are advantages and disadvantages to being the only automator on your team.

Nash · ‎Dec 16 2019

@BrianP wrote: Awesome. Any reason to do the 'Grab device list per network' step ? There is a get all Devices in Org call that is the exact same info as the get list per network. Should save a ton of API calls. https://developer.cisco.com/meraki/api/#/rest/api-endpoints/devices/get-organization-devices GET /organizations/{organizationId}/devices I wanted to check switches/APs set for DHCP against DHCP pools on the MX in their networks. Then if the MX was the DHCP server for that device, check the MX's pool for a reservation for that device's IP. All of my clients have a single network for each physical location, for switches/AP/MX. So once I got inside the network, I could check all of that without having to do a rummage to figure out which MX to compare the switches and APs to. That gave me logic that I could follow more easily. Crunching the returns and writing the spreadsheet slows the process down enough that I didn't hit the call speed limiter.

Nash · ‎Dec 16 2019

I've got a report I run that checks for repeaters. I turn the JSON into objects because I found it easier than manipulating raw JSON later on. It iterates across dozens of orgs, hence my logic below. If you want to see exactly how I did it, look here. It's a work in progress that needs reworking. Order of operations: Get org list Get network list Grab device list per network Convert device list into a list of device objects, since I find it easier (createDeviceObj) Below in report generating chunk, for device in devicelist: check if lanIp is None: (line 572 in script above) If None, check if model is MR If MR, set report value as "Repeater AP?"

Nash · ‎Dec 16 2019

I think Brecht meant, within your loop, you're evaluating against the name of the list. Not the name of the item you're checking. Compare: for item in list: if list == 'thing': print("List matched thing.") Your snippet is doing the same, where item is "network" and list is "networkid" Suggestion: name your list something like "networkIdList" or "network_id_list". Nice and clear what it is, but I suspect there's a Python programmer hissing at me right now.

Nash · ‎Dec 16 2019

Can I suggest that when you're testing your code to make sure a loop works, you run a print statement? In this case, I'd have done (with corrections as per Brecht): networkid = ['ID#1', 'ID#2', 'ID#3'] for network in networkid: print("Checking network in networkid") if network == 'ID#1': print(f"Network: {network} equals ID#1") Then when I run, I see my loop ran and I see if anything matched. For production code, I remove those print statements. Sometimes I have sections of code with print statements after every line if I need to, including within any functions I've written.

Nash · ‎Dec 16 2019

Ah, mea culpa. 🙂 I choose to blame the snow. Good luck with your DNS stuff. This sounds difficult.

About Nash

Nash

Nash King

Community Record

Badges