I will be taking debian.eskimo.com down for about 15 minutes at 9pm this evening to image the machine.
Scientific and Debian are restored to service.
I will be back in the office at approximately 1pm.
The machine which hosts debian and scientific crashed last night. I can’t restore it until around noon due a morning commitment. Please use shellx and eskimo shell servers in the meantime.
Please hold off on Boink until that machine is back up.
We should have a replacement CPU from Intel, which will permanently fix this issue, in about a week.
I don’t know exactly at what time but we will be taking these machines (actually the physical host that hosts them) down for around a half-hour sometime during the evening.
Please use shellx.eskimo.com or eskimo.com while debian and scientific are down.
The machines will be suspended so any running processes will resume when we bring it back up.
This machine has a CPU which has some internal thermal problems, the cores are very hot even though it’s not overclocked and running at the minimum voltage it will run stable and the case temperature is barely above room temperature.
Previously, the cores were hot even when idling, but after adjusting the voltage it idles cool, but under load still gets too hot. This results in throttling (reducing CPU speed to keep temperatures within limits) and as a result less than the performance this machine should provide.
Intel has agreed to replace the CPU but we have to get some information off of it for this to happen and that is what this downtime will be for.
If you haven’t already done so, it’s time to set your clocks ahead one hour, unless you live in Arizona or Hawaii, for Daylight Savings Time (sleep deprivation time).
It was brought to my attention that debian.eskimo.com was not reachable.
The firewall has been fixed, it should be reachable now.
It was also brought to my attention that rup debian.eskimo.com and rup scientific.eskimo.com did not work.
These also have been fixed.
In the News section, you’ll notice a new menu item, “Reviews“.
Click on that link and you’ll find a list of sites that have a listing of Eskimo North with reviews or at least a place for people to write a review.
I really appreciate all of you who have taken the time to take the survey and provide us with some feedback.
I’d like to ask you, if you feel so inclined, click on one or more sites on the Reviews page and take a few moments to write a review of Eskimo North and tell others about us.
Scientific and Debian are back online with much more reasonable core temperatures (they were running 200+ degress). The main issue was CPU voltage. “Auto” on Asus defaulted to 1.38v which was way too high. CPU is stable at 1.05v and MUCH cooler.
Physical id 0: +95.0°F (high = +176.0°F, crit = +208.4°F)
Core 0: +95.0°F (high = +176.0°F, crit = +208.4°F)
Core 1: +89.6°F (high = +176.0°F, crit = +208.4°F)
Core 2: +86.0°F (high = +176.0°F, crit = +208.4°F)
Core 3: +87.8°F (high = +176.0°F, crit = +208.4°F)
I set the CPU voltage too low for the machine to post so it is dead until I can get help to lift it out of the rack, open it up, and reset the BIOS to defaults. As a consequence, scientific and debian are down for the day. Please use shellx or eskimo in the meantime.
My wife and I are going back down this afternoon to try to get this thing working so if you call and nobody answers, please leave a message and if we survive this I’ll call you back when we return.
I am going to take the host for these two servers down for a short period to further attempt to resolve CPU overheating issues. These should be reasonably brief outages as I am going to attempt to adjust the CPU voltage in bios and then reboot, it may take multiple shots at this to find the lowest stable voltage.