prevent >2GB blosc compressors

bin/bootstrax

@@ -15,7 +15,7 @@ How to use
 For more info, see the documentation:
 https://straxen.readthedocs.io/en/latest/bootstrax.html
 """
-__version__ = '2.0.0'
+__version__ = '2.0.1'
 
 import argparse
 import typing
@@ -799,20 +799,20 @@ def infer_records_compressor(rd, datarate, n_fails):
     """
     if n_fails or datarate is None:
         # Cannot infer datarate or failed before, go for fast & safe
-        return 'lz4' if n_fails > 1 else 'blosc'
+        return 'lz4' if n_fails > 1 else 'zstd'
 
     chunk_length = (rd['daq_config']['strax_chunk_overlap'] +
                     rd['daq_config']['strax_chunk_length'])
     chunk_size_mb = datarate*chunk_length
     if datarate < 65:
         # Low data rate, we can do very large compression
-        return 'blosc'
+        return 'zstd'
     if chunk_size_mb > 1800:
         # Extremely large chunks, let's use LZ4 because we know that it
         # can handle this.
         return 'lz4'
     # High datarate and reasonable chunk size.
-    return 'blosc'
+    return 'zstd'
 
 
 ##
@@ -1449,6 +1449,11 @@ def process_run(rd, send_heartbeats=args.production):
                     if not len(md['chunks']):
                         fail("Processing succeeded, but no chunks were written!")
 
+                    if any((chunk['n'] and not chunk['filesize'])
+                           for chunk in md['chunks']):
+                        # E.g. you tried compressing >2 GB chunk using blosc
+                        fail("At least one chunk failed writing!")
+
                     log.info(f"Check that run has ended")
                     rd = get_run(mongo_id = rd['_id'])
                     end_time = rd.get('end')

bin/restrax

@@ -12,7 +12,7 @@ How to use
 For more info, see the documentation:
 https://straxen.readthedocs.io/en/latest/scripts.html
 """
-__version__ = '0.3.0'
+__version__ = '0.3.1'
 
 import argparse
 import logging
@@ -553,6 +553,11 @@ class ReStrax(daq_core.DataBases):
             md_in = storage_backend.get_metadata(dir_in)
             md_out = storage_backend.get_metadata(dir_out)
 
+            if any((chunk['n'] and not chunk['filesize'])
+                   for chunk in md_out['chunks']):
+                # E.g. you tried compressing >2 GB chunk using blosc
+                errors.append(f'For {dir_out}, at least one doc failed to write')
+
             if 'exception' in md_out:
                 errors.append('Writing error!')