Help debug: ssh freezes, ethernet + usb go down [NOT SOLVED]

[MadDenker]

Not much on power consumption for these ethernet jacks, but I did find this:

https://groups.google.com/forum/#!topic ... PGrRTJjReA

They quote power consumption from 500ma - 1.5A! OUCH! I had seen power consumption for the zero to be quoted around 200-300ma, pushing a possible total to as high as 1.8A, so I am willing to bet that power supply is a big issue for many.

After looking over my syslog file, I found the solution posted here, which is working pretty well for now:

viewtopic.php?f=91&t=122601

example of line in /var/log/syslog:
Oct 19 14:00:44 raspberrypi1 rsyslogd-2007: action 'action 17' suspended, next retry is Wed Oct 19 14:02:14 2016 [try http://www.rsyslog.com/e/2007 ]

Come to think of it, I did first run the Zero with a monitor and keyboard, and did not seem to have any difficulties, so that definitely points to the ethernet as the issue.

[sdjf]

I started this topic in the ArchLinuxArm forum because of problems with my Model B, you are running a different model of Pi as well as a different OS, so the error messages are not the same, and fix may be different.

I still have the problem, but very rarely, because I have learned what to do to minimize the chances.

I am still not positive if it is power or the buggey USB driver, but I think that USB can only handle so much going on, so I have learned what kinds of things to not do together, to minimize the risk.

I also run a script 24-7 that does a variety of things for me, among which is it monitors the output of the "free" command because, on my system, with a USB monitor, the amount of used and cached memory are good indicators of when my Pi is a risk of getting into trouble. I have it kill the most memory hungry process at that point, and lose whatever was going on, but at least I don't get stuck having to pull the power plug.

If memory usage is an issue, and it might not be, I would suggest you create a script that loops through creating debugging output, and logging it, say every minute, and after a few outages, you might be able to decipher a pattern.

I think the parameters will be different for different systems and hardware, so my memory limits are likely different than yours.

[MadDenker]

So, I am no longer having this problem. I believe I solved the problem by switching from a long network cable to a shorter one. It was headless anyways, so it didn't matter that I parked it right next to the router with a 4' cable. Everything is the same and it works fine.

I guess it was too much of a voltage drop over the long cable, and the signal became lossy. I hope this helps others trying to troubleshoot this problem.