of course i was unable to reproduce problem yesterday - my workaround was still 
in place! (and working well). doh!

removed my workaround and voilà! problem exhibited again.
to summarise:
 * non-root owner wxsys:wxsys with pidfile in weewx/run/; /etc/init.d/weewx 
changed to reflect this
 * no extensions added, no skin configs, high end RPi (4B 8GB with U-3 SD)
 * tests run while no reports being generated
 * ’sudo /etc/init.d/weewx stop’ and ’sudo /systemctl stop weewx’ fail to stop 
weewx; weewx process running happily but systemctl reports it in failed state

Jul  9 14:45:25 dizzy weewx[621] INFO weewx.reportengine: Copied 5 files to 
/opt/weewx/public_html
Jul  9 14:47:36 dizzy systemd[1]: Stopping LSB: weewx weather system...
Jul  9 14:47:36 dizzy weewx[902]: Stopping weewx weather system: 
weewxstart-stop-daemon: matching only on non-root pidfile 
/opt/weewx/run/weewx.pid is insecure
Jul  9 14:49:37 dizzy weewx[902]: ........................ failed!
Jul  9 14:49:37 dizzy systemd[1]: weewx.service: Control process exited, 
code=exited, status=1/FAILURE
Jul  9 14:49:37 dizzy systemd[1]: weewx.service: Failed with result 'exit-code'.
Jul  9 14:49:37 dizzy systemd[1]: Stopped LSB: weewx weather system.
Jul  9 14:50:16 dizzy weewx[621] INFO weewx.manager: Added record 2020-07-09 
14:50:00 AEST (1594270200) to database 'weewx.sdb'

my interpretation of log:
 * after issuing ‘stop’ command, the command logged “non-root pidfile is 
insecure” then hung for 121 secs
 * weewx carried on and never reported SIGTERM => signal was not sent
 * ‘stop’ command reports failure after timeout and marks state ‘failed’
previous observations:
 * if weewxd process manually killed during command hang, command notices 
weewxd gone and completes successfully

hypothesis: the “non-root pidfile is insecure” msg indicates start-stop-daemon 
refused to send signal to pid found in “insecure” pidfile.
corollary: this would work fine if run as root, not as wxsys:wxsys

Q: is anyone else trying non-root? do they see this behaviour or is it just me?

in the meantime, my workaround goes back in place.
cheers
g-eddy

> On 8 Jul 2020, at 11:38 pm, Graham Eddy <[email protected]> wrote:
> 
> will report in when i can reproduce the problem again...
> thanks
> g-eddy
> 
>> On 8 Jul 2020, at 9:37 pm, mwall <[email protected] 
>> <mailto:[email protected]>> wrote:
>> 
>> when you do '/etc/init.d/weewx stop' or 'systemctl stop weewx', does weewx 
>> eventually stop?
>> 
>> when you do either of these commands, this sends a SIGTERM to weewx, which 
>> tells the main thread to stop.  however, weewx will not actually stop until 
>> all of its threads have completed.  so if you have a report that takes a 
>> long time to run then weewx may not shut down until well after you send the 
>> SIGTERM.
>> 

-- 
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 [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/weewx-user/A8303A88-AC66-4F3A-B424-EB21BF14B53A%40gmail.com.

Reply via email to