Hi all~
I have a question. I found a great script and plan on the web for utility monitoring on the cheap by Greg Sowell. The idea is fantastic. I have implemented this script and hardware on 4 routers. Three of the routers are RB450’s or 493’s running 6.10, and the 4th is a ccr1009 running 6.17. I have scheduled the “scheduler” to run the script every minute. I have also set up the email utility to email me when the “power” goes down.
The three RB450/493’s run the script with out issue for about a week. Then I start getting errors. Every four hours and eleven minutes, several of the interfaces (including the power monitor interface) go down. I get OSPF errors, and traffic stops flowing for about 2 seconds. I will even get an email saying the interface went down. However, I have had no trouble on the ccr 1009.
While troubleshooting the errors, I disabled the scheduler and the errors went away. I re enabled the scripts on a test router and the errors come back after about a week.
I’m stumped trying to figure out what is causing this problem. Has anybody used the the Sowell “Power Monitoring on the Cheap” script? Did it work? Also, has anybody seen this behavior (of a regular timed error) using the scheduler?
I also think it is a great idea, and implemented it.
With limited success.
I am now back in the “lab” running tests, on a 750 (ROS v 6.19). When the power fail all goes well and the script notifies me. When the power comes back the port ‘dies’ and that the status stay at no-link.
All I could do to get the port back is to reboot the router.
I am still looking at a solution.
Is there a ether port ‘reset’ instruction somewhere then I can incorporate that into the script?
Any one else with this difficulty or a better solution?
What is your port settings ? Try 10M / fullduplex / no flow crtl… But not a lucky thing to loop back a port to itself.
Put there a realy cheap old home router, and if you can ping it… there is no power outage.
(If you have serial port on you RB, in /port/stats you can read back the serial port lines. Put relay there and read… )
I use the same script as you to monitor my batteries and its fine (been working for years). The difficulty that I (we) have is t0 get notified (email) when the mains go off. For this Greg Sowell made a cheap (and nasty) but also sometimes working suggestion to monitor MAIN POWER. So that is what this is about.
Thanks, yes I can see your reasoning and it is good.
On the 750 however there is no voltage variable so I can not use that script. I use the script on the 450’s and there I also measure temprature.
I am still running the Greg Sowell suggestion on a test 750 and get mixed results. The most correct result thus far is when there is a power loss. The notification get sent and all is well. It is when the power comes back that the port “fail” - sometimes, not every time… and remain down until reboot. I have tried all different options possible on the port settings but to no reliable result.
The idea is to be warned well ahead of the battries running out that there is a problem with the mains. I have deep discharge protection on all battery packs that will cut all power below a safe voltage. Hence I want to avoid that cut and if possible bring (take to the site) in more power. (my supplies are designed to accept fresh batteries without shutdown)
Your implementation and resoning are actually quite correct, I did not think of that.