Giter Club home page Giter Club logo

Comments (5)

simonw avatar simonw commented on July 22, 2024

Here's the change:

diff --git a/datasette/database.py b/datasette/database.py
index ffe94ea7..e56dcf03 100644
--- a/datasette/database.py
+++ b/datasette/database.py
@@ -85,12 +85,17 @@ class Database:
             return "db"
 
     def connect(self, write=False):
+        kwargs = {
+            "uri": True,
+            "check_same_thread": False,
+        }
+        if write:
+            kwargs["isolation_level"] = "IMMEDIATE"
         if self.memory_name:
             uri = "file:{}?mode=memory&cache=shared".format(self.memory_name)
             conn = sqlite3.connect(
                 uri,
-                uri=True,
-                check_same_thread=False,
+                **kwargs,
             )
             if not write:
                 conn.execute("PRAGMA query_only=1")
@@ -110,9 +115,7 @@ class Database:
             qs = ""
         if self.mode is not None:
             qs = f"?mode={self.mode}"
-        conn = sqlite3.connect(
-            f"file:{self.path}{qs}", uri=True, check_same_thread=False
-        )
+        conn = sqlite3.connect(f"file:{self.path}{qs}", **kwargs)
         self._all_file_connections.append(conn)
         return conn

I don't want to commit this until I can run some kind of load test that demonstrates a problem with the default connection mode that is solved by switching to IMMEDIATE.

from datasette.

simonw avatar simonw commented on July 22, 2024

I tried running this locustfile.py using python -m locust -f locustfile.py and then simulating 100 users.

I ran sqlite-utils enable-wal data.db and then datasette data.db --root and used the /-/create-token page to create a token which I set as the DS_API_KEY environment variable.

from locust import HttpUser, task, between
import random
import json
import os

class APIUser(HttpUser):
    host = "http://localhost:8001"
    # wait_time = between(0.05, 0.1) # in seconds

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.id_counter = 0
        self.api_key = os.environ.get('DS_API_KEY')
        if not self.api_key:
            raise ValueError("DS_API_KEY environment variable is not set")
        self.headers = {
            'Content-Type': 'application/json',
            'Authorization': f'Bearer {self.api_key}'
        }

    @task(3)
    def get_data(self):
        self.client.get("/data", headers=self.headers)

    @task(3)
    def get_mytable(self):
        self.client.get("/data/mytable?_sort_desc=id", headers=self.headers)

    @task(2)
    def get_mytable_with_id(self):
        self.client.get("/data/mytable?id=5", headers=self.headers)

    @task(4)
    def post_create_data(self):
        self.id_counter += 1
        payload = {
            "table": "mytable",
            "rows": [
                {
                    "id": self.id_counter,
                    "name": random.choice(["Tarantula", "Black Widow", "Huntsman", "Wolf Spider", "Jumping Spider"])
                }
            ],
            "pk": "id",
            "replace": True
        }
        self.client.post("/data/-/create", data=json.dumps(payload), headers=self.headers)

On both my Mac and GitHub Codespaces I was unable to get it to trigger any database lock errors. Mac results:

CleanShot 2024-06-19 at 10 54 36@2x

CleanShot 2024-06-19 at 10 54 47@2x

I tried on Codespaces as well because I thought maybe my M2 performed too well and didn't trigger errors.

from datasette.

simonw avatar simonw commented on July 22, 2024

The currently open challenge here is this: create a locustfile.py which reliably triggers database locked errors when exercising a Datasette instance calling the JSON write API (and reading other endpoints at the same time).

from datasette.

simonw avatar simonw commented on July 22, 2024

Here's Codespaces Lotust run with WAL enabled (the macOS one above didn't have WAL enabled):

CleanShot 2024-06-19 at 11 57 33@2x

CleanShot 2024-06-19 at 11 58 39@2x

I used https://github.com/datasette/studio to run this

from datasette.

simonw avatar simonw commented on July 22, 2024

I'm going to push this code anyway, and exercise it in an alpha for a bit.

from datasette.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.