FYI: "sensors" of 10.0 generating false alarms
Hi, added bug #160542 for "sensors" of 10.0 generating false alarms for my nForce4 bases mainboard. Maybe someone can check the current "sensors" package for his/her hardware... Regards, Ulrich
added bug #160542 for "sensors" of 10.0 generating false alarms for my nForce4 bases mainboard. Maybe someone can check the current "sensors" package for his/her hardware...
The sensors package can not be adjusted automatically for every mobo, as the wiring of the "sensor" chips is different each time. In my experience of 2 mobos (one nforce4), sensors default settings were both utterly useless. You will need to make an /etc/sensors.conf for your mobo model. A correct file for a Gigabyte GA-K8NF-9 is on my website. Btw your bug report isn't going to help anyone. lm_sensors expects everyone to edit sensors.conf for their mobo, you obviously haven't done it yet. Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
On 24 Mar 2006 at 19:52, Volker Kuhlmann wrote:
added bug #160542 for "sensors" of 10.0 generating false alarms for my nForce4 bases mainboard. Maybe someone can check the current "sensors" package for his/her hardware...
The sensors package can not be adjusted automatically for every mobo, as the wiring of the "sensor" chips is different each time. In my experience of 2 mobos (one nforce4), sensors default settings were both utterly useless. You will need to make an /etc/sensors.conf for your mobo model. A correct file for a Gigabyte GA-K8NF-9 is on my website.
Btw your bug report isn't going to help anyone. lm_sensors expects everyone to edit sensors.conf for their mobo, you obviously haven't done it yet.
As you pointed out what I have not done, but not where I can read what I should have done, this message doesn't help me much. I don't want to study electrical engineering before I can use some monitoring software. (You do not state where you got your wisdom from, so don't complain if someone else dies stupid) Regards, Ulrich
Well, while I don't feel guilty if someone else dies stupid (each to their own), I can try to give some pointers ;)
As you pointed out what I have not done, but not where I can read what I should have done, this message doesn't help me much. I don't want to study electrical engineering before I can use some monitoring software.
I was pretty annoyed too to have to dive into the guts of lm_sensors, but that's the current state of its development. Run sensors-detect to get info about your hardware. I recommend deleting every section from sensors.conf which isn't relating to your chip, as contrary to sensors's claims, the previous sections spill into the next one. Edit sensors.conf according to the instructions in the package somewhere. Hope that your mobo is very close to one of the sections already in sensors.conf, because otherwise you're at a point where a study of electrical engineering is advantageous. Expect to have to swap labels - if something constantly reads around -12V, the +5V label next to it isn't going to be pot luck. Not all mobos monitor every voltage. You could ask on the lm_sensors mailing list. Once correctly configured, I find even the fan control is quiet good... but it's not automagic to get to that point. You'll need to fix up the normal CPU temperature range in any case, as nobody can know beforehand what chip you'll stick on. I'd venture to say, lm_sensors is currently only for tinkerers unless you hit lucky. Therefore you're probably best to close the bug again, my guess is that SUSE isn't going to be able to bring lm_sensors from pre-school to adulthood. Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
On 24 Mar 2006 at 21:54, Volker Kuhlmann wrote: [...]
Therefore you're probably best to close the bug again, my guess is that SUSE isn't going to be able to bring lm_sensors from pre-school to adulthood.
Considering the amount of harddisk fillup SuSE Linux already has, a few "known to work" sample files of a few Kb each wouldn't be a bad idea IMHO. Maybe we could even get some mainboard vendors to give out information or ready-to-use configuration files. Regards, Ulrich
participants (2)
-
Ulrich Windl
-
Volker Kuhlmann