dear sbf, this issue is about the apparance of the creation of "1970" .csv files (spot table)
this issue has also been reported by snowmiss at 24sep2014 and earlier (the google-code time)
since a time >5weeks, i experience, very regular (each 3--6days), the creation of the "famous" 1970 files
the cause of the appearance might be:
-- during winterperiode it takes longer for the inverter to be in "monitoring" mode (no-feed-in)
i have not yey been able to adapt my scripts to "grap" the std-out when these files are created,
but what i can see - via the information captured after the executions:
- SBFspot did not receive "real inverter time" (this can be found in the stdout.txt)
or to be more accurate - not for the command where SBFspot does get its DATE from
label: current inverter time
- several data fields are missing in the spot data
- the records in the spot table have also the datetime 19700101 0101
note: parameter "CSV_Spot_TimeSource=Inverter"
- observed with version 301 (304 in use since 3 days)
as this issue is still apearing, -> may i suggest to apply the following detection -> workaround
- before the creation of the .csv file -> to validate the date
when the date is 19691231 or 19700101 use then the "system date" for the identification of the files
when the time for the recrod to be inserted in the table is 196912312hmmss or 197001010hmmss
use then the actual "system time" for the record
the error appears less frequent but the effect of the "undesired" output files is a problem,
note - it's not clear what is written to the SQL data in this case
kr wim
this issue has also been reported by snowmiss at 24sep2014 and earlier (the google-code time)
since a time >5weeks, i experience, very regular (each 3--6days), the creation of the "famous" 1970 files
the cause of the appearance might be:
-- during winterperiode it takes longer for the inverter to be in "monitoring" mode (no-feed-in)
i have not yey been able to adapt my scripts to "grap" the std-out when these files are created,
but what i can see - via the information captured after the executions:
- SBFspot did not receive "real inverter time" (this can be found in the stdout.txt)
or to be more accurate - not for the command where SBFspot does get its DATE from
label: current inverter time
- several data fields are missing in the spot data
- the records in the spot table have also the datetime 19700101 0101
note: parameter "CSV_Spot_TimeSource=Inverter"
- observed with version 301 (304 in use since 3 days)
as this issue is still apearing, -> may i suggest to apply the following detection -> workaround
- before the creation of the .csv file -> to validate the date
when the date is 19691231 or 19700101 use then the "system date" for the identification of the files
when the time for the recrod to be inserted in the table is 196912312hmmss or 197001010hmmss
use then the actual "system time" for the record
the error appears less frequent but the effect of the "undesired" output files is a problem,
note - it's not clear what is written to the SQL data in this case
kr wim