New Primenet feature! Email about problem computers

[Prime95]

I've just implemented a new feature to help users monitor their computers' health. There are 2 new email options you can set for each of your computers.

The first option tells the server to email you if the computer is an hour of more late in sending the daily expected completion dates.

The second option tells the server to email you if the computer returns a suspicious LL result. Note that when you first enable this option, the server will email you about all your suspicious LL results since the start of 2014.

To set these new features go to the CPUs web page, http://www.mersenne.org/cpus/ then click on each computer you wish to monitor to enable emailing.


Some notes:

1) It is somewhat annoying to have to turn these features on for each computer, but doing so lets you choose different parameters for each computer. For example, my laptop which sometimes is off for a full day while traveling is set to email me only if it is a day late - my other computers are set to email me immediately.

2) To more frequently monitor computers for crashes, edit prime.txt and set DaysBetweenCheckins less than one. I set it to 0.125 so that I'll know within 4 hours if my computer has gone offline. Doing so will make it impossible to use the Options/Preferences dialog box.


Now if I just had a way to monitor cudalucas, I'd be all set! Maybe if someone wrote a Windows and Linux utility that regularly sent some kind of message to the server if cudalucas was running, then I could implement a similar email feature for cudalucas. Lots of details to ironed out.

[Xyzzy]

Originally Posted by Prime95

The second option tells the server to email you if the computer returns a suspicious LL result.

[swl551]

Originally Posted by Prime95

...
Now if I just had a way to monitor cudalucas, I'd be all set! Maybe if someone wrote a Windows and Linux utility that regularly sent some kind of message to the server if cudalucas was running, then I could implement a similar email feature for cudalucas. Lots of details to ironed out.

I could probably help with a utility that monitors CudaLucas checkpoint and or results files and could submit values over a GET url.

I'd need a lot more info before starting anything.

Let me know.

[Prime95]

I also just changed the results web page to highlight in yellow any unverified result that has more than one clean LL result. This usually happens when a DC result is submitted that does not match the first result.

[ixfd64]

You know, I was just thinking about suggesting this feature the other day. Great minds think alike!

[TheMawn]

What happens during server outrages? Every so often I will see a message indicating a failed daily report and a retry in 70 minutes.

[Prime95]

Originally Posted by TheMawn

What happens during server outrages? Every so often I will see a message indicating a failed daily report and a retry in 70 minutes.

If there is a server outage during the time your computer is supposed to check in, then when the server is restarted there is a significant chance the hourly email job will run before your computer checks in.

Fortunately server outages are infrequent nowadays.

[petrw1]

Originally Posted by Prime95

I've just implemented a new feature to help users monitor their computers' health. There are 2 new email options you can set for each of your computers.

The first option tells the server to email you if the computer is an hour of more late in sending the daily expected completion dates.

The second option tells the server to email you if the computer returns a suspicious LL result.

Thanks.....great