CyberDruid

FAHmon is handy but not completely accurate and total reliance on it for running your Farm may lead to less than optimal performance. If you have a largish Farm of several cores inevitably some sort of issue will occur that stalls the FAH core or shuts down the Core without automatically restarting it. When this occurs the only thing FAHmon is looking at is the Work Log in the FAH folder. Oftentimes there will be more than one log which can confuse FAHmon throwing false warnings that something is wrong with the client. Other glitches may prevent FAHmon from detecting that a core is halted.

FAHmon is not 100% useful for monitoring a Farm. The best way to ensure no drop in production is to monitor every rig via VNC or your favorite remote desktop SW so that you can compare the direct log with the one FAHmon is reading. You need to check up on them to make sure FAHmon is not giving you false info.

On several occasions I found FAHmon reading an old log and reporting a core not working that actually was working. Other times FAHmon reported that a Core was working when in fact it was not. That is the worst sort of error because unless you notice that the percentage and steps are not changing you would let it alone and not know to restart it.

Here is a basically fool proof way of monitoring the progress of your Farm.

If you followed the steps in my other Folding FAQ you will have VNC and the console version of FAH installed on all of your rigs. Once a day you will need to open up the desktop of each rig via VNC at your local computer and send CTRL+ALT+DEL to the machine you want to check. If the performance in Task Manager indicates anything less than 100% CPU usage you will need to open the console(s) on the machine in question. This will restart a stopped core and only temporarily interrupt a running core. You will not lose a WU by doing this. You may inadvertantly start more than one core of FAH running per machine ID. In that case you simply use task manager to stop the additional process; but usually FAH detects that a core is already active and closes by itself.

In most cases you will only need to send CTRL+ALT+DEL and see that the CPU is at 100% and go to the next rig and repeat. Although more time consuming than simply looking at FAHmon it will ensure you lose as few WUs as possible through stalled cores and inaccurate FAHmon reporting.

Hope that helps keep your Farm @ 100% 24/7 for Team 37726

CyberDruid

woop

Thanks....this will come in useful. I've got my eye on a couple of new folding rigs

__________________
An unused CPU cycle is a terrible thing to waste - Join team #37726 and Fold for the cure

FAQ's How to: Clock Skew Can Save Your Overclock and Memory
How To: Reseat your ASUS NB/SB heat pipe system and shave off some C's

CyberDruid

Good to hear! Keep in the Fold.

__________________
I Build Custom Rigs and Work Acrylic. PM for details.
Everything You Ever Wanted to Know About CyberDruid

xxmonkey321xx

I have a folding farm on the rise, too! I use RealVNC Free Edition.. works good.

Anyways, great FAQ. Keep it up! :)

CyberDruid

Great to hear man--it's a good thing to be doing with those clock cycles!

__________________
I Build Custom Rigs and Work Acrylic. PM for details.
Everything You Ever Wanted to Know About CyberDruid

jrabb1920

I have not had a problem with Fahmon reporting inaccurate info so far, but I still check my rigs the old fashion way even though I use fahmon.