bdcsv.py returns "The datetime zone id '00:00' is not recognised"

Question

I'm trying to run bdcsv.py:

$ sudo python /opt/bluedata/bundles/bluedata-epic-entdoc-minimal-release-3.7-2207/scripts/monitoring/bdcsv.py \
     -c localhost \
     -f cred.json \
     -s 2018/02/07-00:00:00 \
     -e 2018/02/07-23:59:59

I received the error below when using my own start and end values, so for this post I used the start and end values from the example in the BlueData docs.

Running the above returns the following error (I've formatted the json to make it more readable):

processing data for virtual node: bluedata-40 ...
error: {  
   "error":{  
      "root_cause":[  
         {  
            "type":"parsing_exception",
            "reason":"[date_histogram] failed to parse field [time_zone]",
            "line":1,
            "col":477
         }
      ],
      "type":"parsing_exception",
      "reason":"[date_histogram] failed to parse field [time_zone]",
      "line":1,
      "col":477,
      "caused_by":{  
         "type":"illegal_argument_exception",
         "reason":"The datetime zone id '00:00' is not recognised"
      }
   },
   "status":400
}

Any idea what is going wrong here?

Answer 1

The same error occurs when running bdusage.py.

The time zone seems to be passed to the ElasticSearch query in the wrong format. In both scripts, you will find the following lines (I added comments inline for clarification):

tz = time.timezone / 3600 * -1           
if tz < 0:
    tzstr = str(tz).zfill(3) + ":00"     # negative tz will produce strings like "-06:00"
else:
    tzstr = str(tz).zfill(2) + ":00"     # positiv tz will return e.g. "01:00"

tzstr will later be included in the query. The error you described appears only if the time difference is >= 0 hours, because ElasticSearch requires the time zone be in a format like +01:00 or -01:00.

Fix it by replacing the last line in above code like this:

tzstr = "+" + str(tz).zfill(2) + ":00"    # positiv tz will now return e.g. "+01:00"