On Tue, Oct 29, 2013 at 8:46 PM, Bastien ROUCARIES <roucaries.bast...@gmail.com> wrote: > On Tue, Oct 29, 2013 at 8:25 PM, Aron Xu <a...@debian.org> wrote: >> On Wed, Oct 30, 2013 at 3:11 AM, Bastien ROUCARIES >> <roucaries.bast...@gmail.com> wrote: >>> On Tue, Oct 29, 2013 at 7:56 PM, Aron Xu <a...@debian.org> wrote: >>>> On Wed, Oct 30, 2013 at 2:40 AM, Bastien ROUCARIES >>>> <roucaries.bast...@gmail.com> wrote: >>>>> On Tue, Oct 29, 2013 at 7:35 PM, Aron Xu <a...@debian.org> wrote: >>>>>> On Wed, Oct 30, 2013 at 1:15 AM, Tollef Fog Heen <tfh...@err.no> wrote: >>>>>>> ]] Aron Xu >>>>>>> >>>>>>>> > IPMI would be lovely, but I'm not sure we can locate a board right >>>>>>>> > now with >>>>>>>> > that - so, we may have to fix remote management with a remotely >>>>>>>> > controlled >>>>>>>> > power/reset box - I believe they exist (something else I've been >>>>>>>> > looking >>>>>>>> > into). If the DSA already use some then I'd be interested to hear >>>>>>>> > which :-) >>>>>>>> >>>>>>>> I don't know if IPMI is available, but there is certain kind of PCI >>>>>>>> device that can help with remotely power on/off the machine controlled >>>>>>>> by SMS. I'm curious if DSA think IPMI is mandatory for buildd and >>>>>>>> porterbox. >>>>>>> >>>>>>> We would very much like «reasonable remote access». Whether that's IPMI >>>>>>> onto a BMC or serial console which can interact with the boot loader and >>>>>>> a network-enabled power strip is less important. Of course, having nice >>>>>>> features like mounting of ISOs over HTTP and such is a nice bonus, but >>>>>>> not a requirement. >>>>>>> >>>>>>> We haven't really talked about how and when it should be enforced, but >>>>>>> I'm reluctant to take on more porter hardware that lacks reasonable >>>>>>> remote management. >>>>>>> >>>>>> >>>>>> If we can find a way of letting Loongson 3A board supports remote >>>>>> console then you are able to re-install the system because PMON have >>>>>> networking support and can boot the system from tftp. Power control >>>>>> can be done by hacking the on-board power button pins. >>>>> >>>>> It could be done trivally from a chip arm card. Using socat from a tty >>>>> to a ssh tunnel >>>>> see http://www.dest-unreach.org/socat/doc/socat-ttyovertcp.txt >>>> >>>> Looks really cool, and I think it's doable to support power control >>>> like what you've suggested already. >>> >>> What are the safety specification appliable by DSA ? Main tension ? >>> Does the board have a power brick ? >>> >> >> I'm not sure about DSA's opinion, and here is the information about >> the board. It is an almost standard ITX one, and we've put it in an >> ITX chassis retired from a ~2006 Lenovo PC, using its power supply. >> The board has some pins for connecting power bottons (Power and >> Reset), though we are not using it because it looks not fit to the >> connector of the chassis. There is a dedicate button on the board to >> power on/off the machine as well. We used the on board button and no >> hard reset needed/conducted since successful installation of hardware. >> The mentioned ITX machine (6100 model) available for purchase is just >> a complete PC box. > > The mini itx does not specify a power connector.... So if you use an > atx power control do something like this > http://www.mupuf.org/blog/2013/05/11/wtrpm-a-web-based-wt-suite-to-power-up-slash-down-your-computers/
Note that I do not recommand to do that this guy has done due to galvanic isolation problem. You could fry your board with something like this! Always use optocoupled MOS, not directly MOSFET > > >> >> Thanks, >> Aron Xu -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAE2SPAbsCCniEmq2p4eDoNGGrUkGFDeY2JSgKh=qwdzbrki...@mail.gmail.com
