Merge pull request #1621 from mabel-dev/#1620

#1620

.github/workflows/regression_suite.yaml

@@ -1,7 +1,9 @@
 name: Regression Suite
 
 on:
-  push
+  push:
+  schedule:
+    - cron: "0 4 * * *"
 
 jobs:
   regression_matrix:

opteryx/__version__.py

@@ -1,4 +1,4 @@
-__build__ = 459
+__build__ = 461
 
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

opteryx/managers/cache/memcached.py

@@ -12,7 +12,8 @@
 """
 This implements an interface to Memcached
 
-If we have 10 failures in a row, stop trying to use the cache.
+If we have 10 failures in a row, stop trying to use the cache. We have some
+scenarios where we assume the remote server is down and stop immediately.
 """
 
 import os
@@ -116,4 +117,8 @@ def get(self, key: bytes) -> Union[bytes, None]:
 
     def set(self, key: bytes, value: bytes) -> None:
         if self._consecutive_failures < MAXIMUM_CONSECUTIVE_FAILURES:
-            self._server.set(key, value)
+            try:
+                self._server.set(key, value)
+            except:
+                # if we fail to set, stop trying
+                self._consecutive_failures = MAXIMUM_CONSECUTIVE_FAILURES

opteryx/utils/file_decoders.py

@@ -160,7 +160,9 @@ def parquet_decoder(buffer, projection: List = None, selection=None, just_schema
         selected_columns = []
 
     # Read the parquet table with the optimized column list and selection filters
-    table = parquet.read_table(stream, columns=selected_columns, pre_buffer=False, filters=_select)
+    table = parquet.read_table(
+        stream, columns=selected_columns, pre_buffer=False, filters=_select, use_threads=False
+    )
     if selection:
         table = filter_records(selection, table)
     if projection == []: