No worries, I need to do a laser starter page. I will run it past you before I go public with it, just in case I am all wrong.
7 Likes
You lasered the image of a pen and the structure of wood? 
2 Likes
Ryan, why is your board not in with the bart boards on the wiki?? Waiting for stock?
1 Like
3 Likes
Post deleted?? Why??
1 Like
It was an accident…I thought I had replied directly to you, and was trying to check because it didn’t look like I did… I restored it as quick as I could lol
1 Like
I had a crazy issue today, I could not connect with my tablet (using AP mode) drove me nuts. Nothing worked.
I just updated chrome browser on it and now everything works again.
If the tablet is not connected to the internet ever, how could anything have changed for it to stop working? That was weird. Troubleshooting stuff like that is going to suck…
Phone and laptop connected fine, tablet would jam it up and need a restart to connect with anything.
1 Like
Some stuff in older Chrome versions have time-outs, so they just stop working for negotiating secure connectuons, even if the other side hasn’t changed. It’s a nightmare, sometimes. (I am the IT department for my company, so I have to deal with it.) Also I have made it known that if you bring your phone, tablet, laptop, or whatever into the office and it connects to the server, I will be … displeased, and there will be consequences. (Though I’m actually usually pretty chill about it. I just won’t support it.) Usually I just avoud problems by limiting to very few people what the actual wifi password is and just letting people connect to the ‘guest’ network which is firewalled from the servers.
4 Likes
Adding a new macro…“~” is resume. I only lost a quarter sheet of MDF to that one.
In my panic I was able to cause another low memory shutdown, opening and refreshing multiple browser windows.
Still lots of learning to do.
Currently, 3.7.6 firmware has a couple bugs, so skip that one. I sent out all the boards with 3.7.5 I think so everyone should be good.
Exciting day.
2 Likes
Okay, also the big batch is finally on its way. The new batch of micro usb boards is not here so I suppose these will mostly be USB-C.
1 Like
LOL. I hear you!!
1 Like
I’ve said it before. Debugging networking is brutal. It is really nice how well it works. Until it doesn’t.
I don’t want to give everyone a complex about it though. It would be nice if there was some sort of support escape hatch when things are really not going well.
3 Likes
I hear that. When I was working at the CS dept in Uni, if a prof ever asked for root access on his machine, it started in writing, and part of the hand over was we gave it to him in person, and he had to change it to something we didn’t know on the spot. We had several faculty members reconsider their requests, and only had to explain it to one poor idiot. But we knew the ones who still took it knew what they were in for.
Now, I’m just thrilled I don’t have to support JS/HTML/CSS shims for IE6.
2 Likes
I doesn’t help when you have a moderately gruntled co-worker who you know can crash all the Cisco routers with a single packet. Crash, as in factory reset time. And I’m not talking about cutesy Open-WRT routers, I’m talking Cisco infrastructure routers. And he knows the vacation schedule of the network teams…
Why keep such a dangerous individual? First, see above paragraph, second, he was part of the security team.
2 Likes
Absolutely. Please tag me in any connectivity issues I miss and I will try to build up a troubleshooting/possible solutions section asap.
3 Likes
Just a humorous side story.
I had a colleague who worked with me and created embedded software that we use for testing spacecraft. After a particularly brutal coding marathon trying to debug an obscure problem, he got his latest feature working on a test platform and deployed it to a machine in one of our labs where we needed that particular test capability for a spacecraft we were creating special avionics and flight software for. He then plugged that machine into the lab network and headed out for a couple of days off.
I got a call at 3AM from the head of the information security office and head of our NOC, who had been troubleshooting a complete outage of the entire corporate network across two campuses. The only thing they could figure out was that a particular machine in a particular lab was doing something they didn’t understand in a way that blasted firewalls, intrusion detection, switches, phones, embedded controllers, etc. all over the place. It wasn’t supposed to be able to do that, and our firewall provider and switch gear provider top tier support both said that it was impossible for that to happen.
In his tired state, my colleague had plugged that test system into the wrong port on a special switch that we had for that test environment, and instead of spewing whatever crap it was doing to a single little lab it had obliterated the enterprise network for the entire evening and into the wee hours.
I gave the OK for their team to sever the connection to that test system.
We re-created that scenario a week or so later with a mini setup like our enterprise network with the various providers all participating to capture traffic created in that little bit of mayhem. All were very interested to understand what we had managed to do to break things that spectacularly.
A side, side note… I ran custom OpenWRT and Alpine Linux based firewalls that I had built for our embedded ground support equipment in that lab. They stayed up, and the embeded hardware behind them was happily continuing to do their thing in their little island of automation throughout the mayhem.
3 Likes
Had a similar sounding situation where I worked almost 2 decades ago. We had some equipment that we had designed where several units were streaming data to one another via UDP. We also had control data being sent over a TCP connection. Originally in smaller deployments and in our testing this was all fine but we ran into an interesting issue once it was deployed at full scale. On startup, each unit would send out an ARP request to figure out the address of their remote data targets and then start streaming data to them. Some of these data targets would never actually send any data back, so it was entirely unidirectional. Everything would be working fine in this state. In a limited number of cases, either due to the design of the switches in the LAN or the level of activity on the LAN, the entry in the routing table of some specific switches would either be pushed off the table or go stale and be invalidated. The switch would then be receiving frames destined for a MAC that wasn’t on the table, so it would default to broadcasting the frame, which would then get re-broadcast via other switches that were having the same issues… Usually that would self-rectify when the target responded but we hadn’t designed it with any such requirement to respond. This would basically just periodically DDOS the entire critical infrastructure LAN of some very, very large installations, starting and stopping seemingly at random. We got called in after the on-site IT department couldn’t figure out what was going on and just knew we had good network/infrastructure expertise. They had no idea we were actually the ones causing the issue, funnily enough.
3 Likes
I was mainly happy I just moved to Identity Management when all of the traffic for Disney Worldwide Services in Burbank/N. Hollywood started getting routed to one closet in the Disney Online offices. The only things that were worse while I was there were the back-to-back unplanned emergency tests of the SoCal NOC restoration procedures. And handy reminders of why, in an Enterprise-class NOC, you have multiple power providers entering the building from different entry points. We actually had the entire C-Suite serving ice-cream sundaes to the IT teams after that bit of fun…
2 Likes
Next time delete cached content, see if that helps.