https://github.com/levicook/goliath-postgres-spike
proof of concept ~ goliath accessing postgres asynchronously
https://github.com/levicook/goliath-postgres-spike
Last synced: 10 months ago
JSON representation
proof of concept ~ goliath accessing postgres asynchronously
- Host: GitHub
- URL: https://github.com/levicook/goliath-postgres-spike
- Owner: levicook
- Created: 2012-03-15T20:13:29.000Z (over 14 years ago)
- Default Branch: master
- Last Pushed: 2012-03-15T21:44:46.000Z (over 14 years ago)
- Last Synced: 2024-04-15T09:13:16.878Z (over 2 years ago)
- Language: Ruby
- Homepage:
- Size: 93.8 KB
- Stars: 7
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
This project is a proof of concept designed to confirm the viability of accessing [PostreSQL][1] from [Goliath][2], in a non-blocking fashion.
### Goals Met:
1. Make sure that blocking database operations, like "select pg_sleep(1)", don't interfere with our ability to accept and respond to http requests.
1. Limit, without blocking, the number of database connections we're holding open.
1. Don't crash in a giant ball of fire under minor load.
### Running the demo server
1. Make sure you have postgres running locally
1. Make sure you have ruby 1.9.x and bundler installed
1. Clone the project, bundle, etc.
1. Launch the server
`$ ./api --port=9000 --verbose`
`[75693:INFO] 2012-03-15 15:23:06 :: Starting server on 0.0.0.0:9000 in development mode. Watch out for stones`
1. Confirm it's basically working
`$ time curl -i localhost:9000`
`HTTP/1.1 200 OK`
`Content-Type: text/plain`
`Content-Length: 8`
`Server: Goliath`
`Date: Thu, 15 Mar 2012 21:28:12 GMT`
`pg_sleep`
`real 0m1.025s`
`user 0m0.002s`
`sys 0m0.003s`
### Resources:
- [PostgreSQL][1]
- [Goliath][2] -- Rubyists should really spend more time with this project.
- [EM::Synchrony][3] -- Because it makes node.js kids jealous.
- [EM::Postgres][4] -- I was especially glad to run into this project, as it's exactly
what I was expecting to have to implement on my own.
- [OldMoe][5] -- Dated, but very interesting work on non-blocking database IO. Really highlights the fundamental reasons for pursuing this architecture.
- [StackOverflow][6] -- Because I had the same basic question.
[1]: http://www.postgresql.org
[2]: http://postrank-labs.github.com/goliath
[3]: https://github.com/igrigorik/em-synchrony
[4]: https://github.com/jtoy/em-postgres
[5]: http://oldmoe.blogspot.com/2008/07/faster-io-for-ruby-with-postgres.html
[6]: http://stackoverflow.com/questions/5893524/using-the-postgresql-gem-async
### Results
#----------------------------------------------------------------------
$ ab -d -S -n30 -c5 http://localhost:9000/
This is ApacheBench, Version 2.3 <$Revision: 655654 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/
Benchmarking localhost (be patient).....done
Server Software: Goliath
Server Hostname: localhost
Server Port: 9000
Document Path: /
Document Length: 8 bytes
Concurrency Level: 5
Time taken for tests: 6.045 seconds
Complete requests: 30
Failed requests: 0
Write errors: 0
Total transferred: 3780 bytes
HTML transferred: 240 bytes
Requests per second: 4.96 [#/sec] (mean)
Time per request: 1007.507 [ms] (mean)
Time per request: 201.501 [ms] (mean, across all concurrent requests)
Transfer rate: 0.61 [Kbytes/sec] received
Connection Times (ms)
min avg max
Connect: 0 0 0
Processing: 1003 1006 1015
Total: 1003 1006 1015
#----------------------------------------------------------------------
$ ab -d -S -n30 -c10 http://localhost:9000/
This is ApacheBench, Version 2.3 <$Revision: 655654 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/
Benchmarking localhost (be patient).....done
Server Software: Goliath
Server Hostname: localhost
Server Port: 9000
Document Path: /
Document Length: 8 bytes
Concurrency Level: 10
Time taken for tests: 6.022 seconds
Complete requests: 30
Failed requests: 0
Write errors: 0
Total transferred: 3780 bytes
HTML transferred: 240 bytes
Requests per second: 4.98 [#/sec] (mean)
Time per request: 2007.397 [ms] (mean)
Time per request: 200.740 [ms] (mean, across all concurrent requests)
Transfer rate: 0.61 [Kbytes/sec] received
Connection Times (ms)
min avg max
Connect: 0 0 0
Processing: 1004 1838 2020
Total: 1004 1838 2020