Sept 15th & 16th, 2013

Temperatures high (and the new server with 5 fans blowing hot air in the room doesn't help). Try changing the ventilation set-up and watch temperatures.

Well, it is not very surprising that the room heats-up, is it ? Back in 2009 we had 9 machines each one burning about ~300 Watts, which gives 3.412*2700 ~ 9200 BTUs. Which means that the 9000 BTU unit could possibly manage to keep the room cool. Then we added 4 GPUs on the old nodes, each at ~200W. Then we got n0009 and n0010 with nominal requirements of ~600W each. This got us to ~16000 BTUs at full load. Now, we added the IBM server which should be at about ~1200 Watts. We are now at 5900 Watts, or, ~20000 BTUs and the A/C wouldn't stand a chance at full load (and I have forgotten to add the server's card + the quadro card + the orion machine, make them another ~600W). At the time of writing, four machines are idle, which means that we still have 5900-1400=4500W, or ~15000 BTUs to worry about. No wonder that we can't keep the room cool.

The change of the ventilation combination helped, but this is short term : when the A/C units are turned back to heating mode, we will be in serious trouble.

⇒ Well, it helped, but there are other things at play as well. There is a strong periodicity of the observed temperatures and this is possibly connected with the A/C unit of the outside terminal room failing. See the periodicity of the spikes in the temperature graph :

On the verge of the big meltdown ?

No, the temperature climbed (predictably) till about 2200 hours, then the automatic restart of the outside A/C kicked-in, and temperatures fell back to normal (10 degrees down !) by 2300 hours. But this definitely needs fixing. See how strong the periodicity is after the change in ventilation :

Would it make sense to completely remove the roof panels from above the cluster room ????? ⇒ No, the outside air is humid and will create more problems. Trapped ?

⇒ Try again isolating the room from the external A/C → No, temps went up.

⇒ Last ditch effort : open everything up including roof panels. Let's see if (when ?) the A/C will collect ice.