Did my M1XEP just die?

[IVB]

WooHoo, baby steps!
 
Using a USB->Serial cable that I had I was able to *barely* reach the Elk with just the one straight through cable. (who the hell has crossover cables? What was that pete said about glitch?)
 
I can connect via serial so the main panel is fine. Phew. But i'm now suspecting the M1XEP really is bad, as putting in a legit jumper the lights aren't 100% with what the hard reset claims it should be.
 
Unit is nearly 10 years old, so I s'pose MTBF has to set in sometime.

 

[IVB]

Well, you can check the cable with a multi-meter.  It should be a straight cable, and I believe Elk's serial port uses only pins 2,3 and 5 with a 115,200 baud rate.
 
I used a serial-to-USB FTDI cable with no issues as well.

 
Was using my multi-meter to check it, was sitting upstairs in the nook near the kitchen as its bright not dingy dark. Wife/kids were a little concerned as I kept muttering "huh? *another* crossover cable? why the hell did I ever need this many crossovers".
 
Then again this basement purge would be really funny and cathartic if I didn't look at the pile heading to the electronics recycler and mentally calculate costs. I stopped counting once I realized I was over $1K in "seemed like a good idea at the time"

 

[wkearney99]

..."seemed like a good idea at the time"

I'm out of my boat right now, surrounded by seemed like a good idea at the time!

https://drive.google.com/file/d/0B-Hd0lSE7isudG5Da0pHa3RMaU0/view?usp=drivesdk

 

[MikeB]

Hi IVB -
 
I was able to resurrect an M1XEP that became unresponsive due to a failed firmware update.  PM me if you want more info.

 

[IVB]

WooHoo, thanks to Mike's help I was able to resurrect it! Took several attempts, plus another 5-10 JP2 shorts, and was about to walk away dejected at having to dump $200 just to replace what I have so I did one last full cycle of attempts.
 
BTW one quick note: ElkRP "find" will only work if your M1XEP is on the same subnet as ElkRP. I have 2 subnets as one of my switches is a managed switch (which is how I could tell the M1XEP was transmitting).  I want the M1XEP on a different subnet than the servers as I separate video streaming traffic from low bandwidth/home automation stuff, but yesterday I moved the patch panel cord so they were both on the same subnet. Once I got it all working I moved it back to the desired subnet, then popped into my router config page to determine its IP address and manually typed that into ElkRP.
 
Thanks Mike, I owe you many beers.

 

[wkearney99]

I have 2 subnets as one of my switches is a managed switch (which is how I could tell the M1XEP was transmitting).  I want the M1XEP on a different subnet than the servers as I separate video streaming traffic from low bandwidth/home automation stuff, 

 
Technically you do not "have to" use separate subnets to isolate traffic.  Yes, doing so will pretty much guarantee the traffic won't interfere.  But if you've got decent fast switches arranged "like" subnets you can often get the same results.  
 
The key is making sure the switches that are going to be handling the heaviest amounts of traffic are 'under subscribed'.  That is, using a 24 port GigE switch but only using 12 ports on it.  Assuming, of course, that the 24 port switch has a large enough capacity internally to handle all 24 ports.  Using less than the max gives you some degree of 'room'.  Then chain that into a likewise beefy central switch, with an uplink to your ISP router. 
 
What you're looking to avoid is cross-uplink traffic.  Keep as much as possible within a given switch and not traveling across the uplinks.  That and you really do not want to set up too many levels of nested switches.  I use a 'no more than three hops' rule of thumb.  That was a requirement back in the days of passive hubs, but it's still not a bad practice to continue.

Where this can get gnarly is if there's multi-cast IP traffic being used.  But if you've got that much going on then separate subnets is REALLY going to be a lot of trouble to configure.  VLANs, likewise, can help but have their own set of headaches to consider.
 
Another factor to consider is IP address range size.  Lots of folks default to using a class C range, which supports 253 devices.  Going to a Class B changes that but there have been residential/low-end devices that are sometimes have issues with non Class C range configurations.  As you start getting more devices you start running up against the limits of the ARP tables in the switches.  This is another reason to under-subscribe switches.  With a lot of devices you could eventually run into ARP table woes, but it's pretty unlikely in most residential situations.  Subnet separation would help, but again raises complications with multi-cast and devices that just don't "play well" across subnets.

 

[wkearney99]

WooHoo, thanks to Mike's help I was able to resurrect it! Took several attempts, plus another 5-10 JP2 shorts, and was about to walk away dejected at having to dump $200 just to replace what I have so I did one last full cycle of attempts.

Thanks Mike, I owe you many beers.

 
Wow, that's great news!  
 
Oh, and the initial reason I posted that switch message was to say you can also use a passive hub.  I keep one around just for that purpose.  Plug one in-between the questionably behaving devices along with a PC running Wireshark.  That'll let you listen-in on any traffic going through the whole hub.  
 
Managed switches are great, but pricey.  That and setting up the same sort of thing with a managed switch always seems to take a lot more time/effort.  If only because of having to drag out the manual or scour the web for the right config changes to make it happen.  With a dumb hub you're pretty much plug-and-go.

 

[pete_c]

Good news IVB!

Wrote a bunch next here then just deleted it as my post got mangled.

I do keep main network at 1/2 of a class C to make it easier to manage.

 

[vc1234]

Good news IVB!

Wrote a bunch next here then just deleted it as my post got mangled.

I do keep main network at 1/2 of a class C to make it easier to manage.

Never bothered with anything but class A.  Since about 20 years ago, I've used 10.x.x.x./255.0.0.0 for all my ip gadgets.
 
Traffic separation happens at the switch level (e.g. IP cameras have their own switch).
 
I do have a guest wifi of class C.

 

[IVB]

TBH i've never done anything with separation beyond basic switches, but the ubiquiti Edgeswitch created its own subnet so I just went with it.
 
I don't have anything nested, all the CatX wiring is homerun back to the panel. Just 2x24p switches right now, both run into the router. I only have a 1650sqft house, so hopefully that'll be fine.
 
The one thing I need to think about is APs. I have 3 ubiquiti APs to assure coverage, but right now they're going into the high bandwidth switch. But its too early in the morning to think about that, sabbatical or not.

 

[pete_c]

Here just added more cables and patch panels to the central wiring location. 
 
I have used cat5x for more than networking over the years.  Initially it was my 1-wire network and RS-232 long runs.
 
Also used the cat5e for my Russound keypads, Omnitouch 5.7/5.7e screens. 
 
Router wise started with just two ports; then over the years added multiple LANs / WAN connections more to tinker with.
 
Relative to automation have kept to the wires with no automation running on wireless stuff today.
 
I like to tinker with new automation widgets.  I have always tested these though on their own network and connection to the wan rather than using VLANs.  Firewall nics do VLANs today though.
 
Television stuff started with just RG-6 (initially had satellite / cable) then added multiple runs of cat5e which I use more today.