jennydiep/zotflix

CS 122B: Fabflix

• General

  • Team#: 82

  • Names: Jenny Diep

  • Project 2 LATE Video Demo Link: https://youtu.be/U0qVCBLomhs

  • Project 5 Video Demo Link: https://youtu.be/gSopbQXfAmE

  • Instruction of deployment:

    If you do not have USER mytestuser setup in MySQL, follow the below steps to create it:

    • login to mysql as a root user

      local> mysql -u root -p
      

    • create a test user and grant privileges:

      mysql> CREATE USER 'mytestuser'@'localhost' IDENTIFIED BY 'mypassword';
      mysql> GRANT ALL PRIVILEGES ON * . * TO 'mytestuser'@'localhost';
      mysql> quit;
      

    • prepare the database moviedb

      mysql -u mytestuser-p < create_table.sql
      

    To run this example on local machine:

      1. Clone this repository 
      2. Open IntelliJ -> Import Project -> Choose the project you just cloned (The root path must contain the pom.xml!) -> Choose Import project from external model -> choose Maven -> Click on Finish -> The IntelliJ will load automatically
      3. For "Root Directory", right click "cs122b-spring20-project1-api-example" -> Mark Directory as -> sources root
      4. In `WebContent/META-INF/context.xml`, make sure the mysql username is `mytestuser` and password is `mypassword`.
      5. Also make sure you have the `moviedb` database.
      6. To run the example, follow the instructions in [canvas](https://canvas.eee.uci.edu/courses/26486/pages/intellij-idea-tomcat-configuration)
    

• Connection Pooling

  • Include the filename/path of all code/configuration files in GitHub of using JDBC Connection Pooling.

    context.xml

    DashBoardServlet

    LoginServlet

    ItemsServlet

    PaymentServlet

    AdvancedSearchServlet

    SingleMovieServlet

    SingleStarServlet

    AddMovieServlet

    GenreServlet

    SearchServlet

    MetaDataServlet

    StarsServlet

  • Explain how Connection Pooling is utilized in the Fabflix code.

    maxTotal="100" maxIdle="30" maxWaitMillis="10000"
    

    maxTotal represents the max amount of connections to the db which is 100 connects, maxIdle represents the minimum number of connections and maxWaitMillis represents the amount of time waiting on the connection before timing out.

    Since there are many calls to the db in Fabflix connection pooling is a way to make calls from and to the database more efficient. Instead of establishing new connections for example in each search call, instead it will use a connection that was already established in the pool of connections.

  • Explain how Connection Pooling works with two backend SQL.

    Connection pooling with two backend databases works similar to single database where instead of establishing new connections it chooses existing connections, however the pools are shared between two databases which speeds up queries since there can be multiple queries at once instead of waiting for one query to finish.

• Master/Slave

  • Include the filename/path of all code/configuration files in GitHub of routing queries to Master/Slave SQL.

  • config file: context.xml

  • master only:
    AddMovieServlet
    DashBoardServlet

  • master/slave:
    LoginServlet
    ItemsServlet
    PaymentServlet
    AdvancedSearchServlet
    SingleMovieServlet
    SingleStarServlet
    GenreServlet
    SearchServlet
    MetaDataServlet
    StarsServlet

  • How read/write requests were routed to Master/Slave SQL?

    Read request chooses SQL instance that is linked to its session. The SQL instance that is chosen by the session is picked at random (either the slave or master instance, whatever is localhost) while write requests are always from the master instance.

• JMeter TS/TJ Time Logs

  • NOTE: I did not finish xml implementation so I am using my own query_load.txt file with movies that are only in the original database

  • another note: single aws instance on 8080 is the same as the load balancer instance.

  • Instructions of how to use the log_processing.* script to process the JMeter logs.

    log_processing.py prints out the average TS and TJ time

    Go to set folder ex: .../log/1

    Run:

    python3 ../log_processing.py       
    

• JMeter TS/TJ Time Measurement Report

Single-instance Version Test Plan	Graph Results Screenshot	Average Query Time(ms)	Average Search Servlet Time(ms)	Average JDBC Time(ms)	Analysis
Case 1: HTTP/1 thread	1	231	147.11685	146.91967	Having only one user means there is no wait time for other users to finish their search query thus explains its speed.
Case 2: HTTP/10 threads	2	1649	1481.37576	1481.21719	naturally with only one databases, when adding more users the queries get slower speeds since they have to wait for each query to finish.
Case 3: HTTPS/10 threads	3	1794	1482.13113	1481.89844	I hypothesized that https would be slower since it takes more time due to encryption that https has which seems to be true.
Case 4: HTTP/10 threads/No connection pooling	4	1572	1426.65872	1426.47031	I hypothesized that no connection pooling would have a higher average however I assume connection pooling isn't much faster in single system since you still have to wait for a whole query to finish and that establishing new connections does not cost as much.

Scaled Version Test Plan	Graph Results Screenshot	Average Query Time(ms)	Average Search Servlet Time(ms)	Average JDBC Time(ms)	Analysis
Case 1: HTTP/1 thread	5	230	143.42437	143.17355	As expected to have the same time as no single instance since if there is only one user tied to one session there are not too many calls to overload the database
Case 2: HTTP/10 threads	6	794	696.48106	696.29226	Was expected to be much quicker but was not expected it to be 3x as fast. The logic behind the scaled version being faster is that reads are much faster since calls to the database are offloaded onto the other server.
Case 3: HTTP/10 threads/No connection pooling		928	717.43103	717.27264	Here connection pooling is shown as slower which I hypothesized because without connection pooling the times to create new connections and destroy connections added up for extra time.

substring matching

title: any title that starts with "x"
director: has "x" in it
star: has "x" in it