Project

General

Profile

Bug #2186

Email notification stops working

Added by Anonymous over 4 years ago. Updated over 4 years ago.

Status:
Closed
Priority:
Normal
Assignee:
-
Category:
-
Start date:
05/19/2015
Due date:
% Done:

0%


Description

Occasionally, email notifications stop working
Using git version from 12 Feb 2015
I cannot identify the conditions, I attach my epgsearch.conf - this time, notifications stopped 2 weeks ago
This seems to line up with last change to epgsearch.conf
I use vdradmin to maintain the epgsearch config (3.6.9)
New timers are still being created, just no emails.
Using log level 2, events are logged, but no indication why email notifications are stopped
pendingnotifications.conf is empty

epgsearch.conf Magnifier (34.1 KB) Anonymous, 05/19/2015 01:36 AM

History

#1 Updated by winni over 4 years ago

  • Status changed from New to Assigned

if possible please check your email account settings in the OSD (or with the remote screen of vdradmin) plugin setup and try the "test" function

#2 Updated by keynet over 4 years ago

Yes, I tried the test function, and it works
I think it's something to do with either the number of searches or the format of the search, as I've had this issue before, and it went away when I changed some searches, but I don't see a pattern. Do you have any tool to validate the attached conf file, as that's the thing that's changed?

#3 Updated by winni over 4 years ago

Your epgsearch.conf looks fine, but anyway the problem should not be related to the search timers themselves.
Please try to set the log level of epgearch to debugging (with "-v 3") and check epgsearch.log.

#4 Updated by keynet over 4 years ago

So I added 1 more timer yesterday and now it's working again!
I will revert to the attached file and see if it stops again, let you know

#5 Updated by keynet over 4 years ago

Since you asked me to add the debug flag it's working again, even with the epgsearch.conf that was not working before.
But this is the usual scenario - email stops working for no apparent reason, starts working again for no apparent reason. I restart vdr on a cron job monthly, but also more often for for other reasons, and when epgsearch email stopped working before, it does not start working again after vdr restart - I don't believe it's related to vdr restarts. Are there any other files that "keep state" that I need to check if/when it happens again?

#6 Updated by winni over 4 years ago

"...it does not start working again after vdr restart..."

you only get new notifications if there is new content. So a restart will not produce the same email again.
epgsearch stores all pending notifications about events or timers in pendingnotifications.conf.

#7 Updated by keynet over 4 years ago

I know I only get notifications with new events. What I meant to say is that when it stops working, restarting vdr doesn't fix it, still not working days later.
So now it's stopped working again - first of the month, 04:00 VDR was auto-restarted.
Below is the log. Seems the sendEmail.pl script is no longer returning success after restart.
I checked, there's no stuck sendemail.pl script.
According to the mailserver log, there was no attempt to send anything at 07:48 on 1 June.
So it looks like either the script doesn't run, or is faulty somehow.
What else to look at?
Or just not use the script?

Mon 01.06.2015 01:48:11: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 0 new / 0 modified / 0 deleted timers / 1 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Mon 01.06.2015 01:48:11: mail cmd result: Email was sent successfully!
Mon 01.06.2015 07:48:41: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 2 new / 0 modified / 0 deleted timers / 8 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Tue 02.06.2015 07:48:37: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 3 new / 0 modified / 0 deleted timers / 10 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Wed 03.06.2015 07:48:50: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 2 new / 1 modified / 0 deleted timers / 18 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Thu 04.06.2015 01:23:53: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 0 new / 0 modified / 1 deleted timers / 0 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Thu 04.06.2015 07:48:38: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 3 new / 2 modified / 1 deleted timers / 13 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp
Thu 04.06.2015 12:48:11: starting mail script: sendEmail.pl with parameters: -f "VDR <>" -t -s localhost -u "[epgsearch] update info: 1 new / 0 modified / 0 deleted timers / 0 new events " -o message-charset=UTF-8 -o message-file=/etc/vdr/plugins/epgsearch/epgsearchmail.temp

#8 Updated by winni over 4 years ago

ok, according to your logfile epgsearch try to send emails, but the script sendEmail.pl does not work always. So it's just a problem with sending the emails. Try to use sendmail instead (has to be configured and selected in the plugin setup) or try to contact the author of script.

#9 Updated by keynet over 4 years ago

OK, pretty sure I fixed it.

The difference was the environment under cron and normal / root users, even though the user who starts vdr is always "vdr"
Cron has a very limited environment, and the email script is in /usr/local/bin (FHS and epgsearch (4) recommendation), so not available in the path unless manually invoking it.

SHELL=/bin/sh
MAILTO=root
USER=vdr
PATH=/usr/bin:/bin
PWD=/var/spool/cron
SHLVL=1
HOME=/home/vdr
LOGNAME=vdr

So I used the -m option to specifically point epgsearch to the location, and now it survives cron related restarts.
Thanks for your help. Might be worth noting this detail in epgsearch(4)?

#10 Updated by winni over 4 years ago

  • Status changed from Assigned to Closed

Also available in: Atom PDF