This saga really does sound like what happened to me running wheezy on RPiB - just started doing random stoppages - usually locking me out of ssh at the same time renderingdebugging difficult that I switched to rpi3 and still - touch wood - not failed since. I am sure raspberry/debian introduced a sw change of some kind that affected the older rpi's in some manner. Hope you eventually get to the root cause of this. As I now do not trust my old rpiB am not sure what to do with it - try it on another project or bin it. It's been ticking away doing nothing now ever since I decomiisioned it without failing!! If I was you I would try weewx on the rpi3 and simulator on the rpi2 rather than what you are proposing - even though that proposal is counter-intuitive and not good practice.
On Wednesday, 28 December 2016 04:51:46 UTC+2, Marc Aubry wrote: > done. > i'm running a screen bash weewx session and monitor it until it stop. > thanks. > > My config is pretty standard for rpi2. > i install a raspbian image from raspberry pi foundation website i install > tight vnc server (not needed if i had done the jessie upgrade before) and > then upgrade to the last jessie release. > after that i just follow weewx doc and wiki to run a sofaskin on local > webserver in a ram disk to avoid overwriting on mys sd card and remote ftp > server. external mysql (for the same reason) and forecast extension. > for the syslog i did try to syslog to my ubuntu linux server as describe > in weewx wiki to save writing on the rpi sd card but came back to original > syslog config because it was apparently still writing into the local syslog > file in top of the remote syslog... > i had weewx stop reporting randomly from the beginning, before i did > change the skin or forecast extension or sys log modification.... > i did change the rpi2 board with the same model. > i will now setup a dev config with a rpi3 board and the simulator as > weather station. > right now, i keep this config running like that just if it is useful for > the community to help debugging the software if it's hardware related i > don't see the point to maintain a buggy system and i'll keep it up running > until i can validate my new rpi3 dev platform to switch it in prod platform. > > > > On Wednesday, December 28, 2016 at 9:14:26 AM UTC+7, Marc Aubry wrote: >> >> hi >> do i have to rename it wh23xx.py in /usr/share/weewx/user >> or i just copy it like that. >> i guess that i have to rename it :) >> >> >> >> On Wednesday, December 28, 2016 at 8:26:32 AM UTC+7, mwall wrote: >>> >>> marc, >>> >>> the problem is that pyusb/libusb should timeout when a write/read fails, >>> but that is not happening. but if the write/read is not failing, why are >>> we suddenly not getting any logging? >>> >>> this really smells like a hardware problem, or perhaps something with >>> your system's logging configuration. >>> >>> try the attached wh23xx-0.8p1.py >>> >>> it does non-debug logging before and after every interrupt write to the >>> usb. >>> >>> m >>> >>> -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.