Updated the reload and shutdown mecanism to avoid a loop when shutting down from inside the game. Made sure to have server sync correctly with portal at @reload (some session info were lost before). Some other cleanups.

src/commands/default/system.py

@@ -6,6 +6,7 @@ System commands
 
 import traceback
 import os, datetime, time
+from sys import getsizeof
 import django, twisted
 
 from django.conf import settings
@@ -16,9 +17,10 @@ from src.players.models import PlayerDB
 from src.utils import logger, utils, gametime, create
 from src.commands.default.muxcommand import MuxCommand
 
+# delayed imports
 _resource = None
 _idmapper = None
-
+_attribute_cache = None
 
 # limit symbol import for API
 __all__ = ("CmdReload", "CmdReset", "CmdShutdown", "CmdPy",
@@ -537,12 +539,31 @@ class CmdTime(MuxCommand):
 
 class CmdServerLoad(MuxCommand):
     """
-    server load statistics
+    server load and memory statistics
 
     Usage:
        @serverload
 
-    Show server load statistics in a table.
+    This command shows server load statistics and dynamic memory
+    usage.
+
+    Some Important statistics in the table:
+
+    {wServer load{n is an average of processor usage. It's usually
+    between 0 (no usage) and 1 (100% usage), but may also be
+    temporarily higher if your computer has multiple CPU cores.
+
+    The {wResident/Virtual memory{n displays the total memory used by
+    the server process.
+
+    Evennia {wcaches{n all retrieved database entities when they are
+    loaded by use of the idmapper functionality. This allows Evennia
+    to maintain the same instances of an entity and allowing
+    non-persistent storage schemes. The total amount of cached objects
+    are displayed plus a breakdown of database object types. Finally,
+    {wAttributes{n are cached on-demand for speed. The total amount of
+    memory used for this type of cache is also displayed.
+
     """
     key = "@server"
     aliases = ["@serverload", "@serverprocess"]
@@ -559,11 +580,13 @@ class CmdServerLoad(MuxCommand):
         if not utils.host_os_is('posix'):
             string = "Process listings are only available under Linux/Unix."
         else:
-            global _resource, _idmapper
+            global _resource, _idmapper, _attribute_cache
             if not _resource:
                 import resource as _resource
             if not _idmapper:
                 from src.utils.idmapper import base as _idmapper
+            if not _attribute_cache:
+                from src.typeclasses.models import _ATTRIBUTE_CACHE as _attribute_cache
 
             import resource
             loadavg = os.getloadavg()
@@ -622,10 +645,10 @@ class CmdServerLoad(MuxCommand):
             for row in ftable:
                 string += "\n " + "{w%s{n" % row[0] + "".join(row[1:])
 
-            # cache size
+            # object cache size
             cachedict = _idmapper.cache_size()
             totcache = cachedict["_total"]
-            string += "\n{w Object cache usage: %5.2f MB (%i items){n" % (totcache[1], totcache[0])
+            string += "\n{w Database entity (idmapper) cache usage:{n %5.2f MB (%i items)" % (totcache[1], totcache[0])
             sorted_cache = sorted([(key, tup[0], tup[1]) for key, tup in cachedict.items() if key !="_total" and tup[0] > 0],
                                     key=lambda tup: tup[2], reverse=True)
             table = [[tup[0] for tup in sorted_cache],
@@ -634,6 +657,10 @@ class CmdServerLoad(MuxCommand):
             ftable = utils.format_table(table, 5)
             for row in ftable:
                 string += "\n  " + row[0] + row[1] + row[2]
+            # attribute cache
+            size = sum([sum([getsizeof(obj) for obj in dic.values()]) for dic in _attribute_cache.values()])/1024.0
+            count = sum([len(dic) for dic in _attribute_cache.values()])
+            string += "\n{w On-entity Attribute cache usage:{n %5.2f MB (%i items)" % (size, count)
         caller.msg(string)
 
 # class CmdPs(MuxCommand):