Merge pull request #31 from jas4711/logfiles-live-data-missed

Discuss live data missed.
jim-easterbrook · Oct 31, 2015 · 54e41d1 · 54e41d1
2 parents 8b20e9d + 65335f4
commit 54e41d1
Show file tree

Hide file tree

Showing 2 changed files with 24 additions and 1 deletion.
diff --git a/src/contributors/contributors.txt b/src/contributors/contributors.txt
@@ -7,6 +7,7 @@ x2q
 Robin Kearney                   robin@kearney.co.uk
 Rod Persky
 Morten Høybye Frederiksen       morten@mfd-consult.dk
+Simon Josefsson                 simon@josefsson.org
 
 Translators
 -----------
@@ -20,4 +21,4 @@ Kyle Gordon                     kyle@lodge.glasgownet.com>
 Πέτρος                          nouvakis@sch.gr
 Ramiro                          ramiro.sanchez@telefonica.net
 Rick Sulman                     rick@sulman.org
-Pyttsen                         weather@spacelab.se
+Pyttsen                         weather@spacelab.se
diff --git a/src/doc/guides/logfiles.rst b/src/doc/guides/logfiles.rst
@@ -168,6 +168,28 @@ The raw rainfall data from the outside sensors is the total number of times the
 This number should only ever increase, so the :py:mod:`pywws.Process` module warns of any decrease in the value as it may indicate corrupted data that needs manually correcting.
 The logging message includes the UTC time stamp of the problem data to help you find it.
 
+Live data missed
+----------------
+
+.. code-block:: none
+   :linenos:
+
+   2015-10-30 04:48:19:pywws.ToService(underground_rf):1 records sent
+   2015-10-30 04:49:07:pywws.ToService(underground_rf):1 records sent
+   2015-10-30 04:49:56:pywws.weather_station:live_data missed
+   2015-10-30 04:50:44:pywws.ToService(underground_rf):1 records sent
+   2015-10-30 04:51:31:pywws.ToService(underground_rf):1 records sent
+
+Line 3 indicate that pywws failed to capture live data.
+
+There are two possible causes. One is that a new data record is identical to the previous one so pywws doesn't detect a change. This is unlikely to happen if you are receiving wind data properly.
+
+The more likely reason is that processing the previous record took so long that the next one arrived when pywws wasn't ready for it. "Processing" can include uploading to the Internet which is often prone to delays.  A solution to this is to set "asynchronous" to True in weather.ini. This uses a separate thread to do the uploading.
+
+You may run with higher verbosity to get more information.  The "pause" values should indicate how soon it's ready for the next data.
+
+Note that this is just an occasional missing "live" record though, so if it does not happen often you shouldn't worry too much about it.
+
 "Live log" synchronisation
 --------------------------