nameko-pymemcache 0.1.0

Memcached dependency for nameko services with consistent hashing

nameko-pymemcache

Memcached dependency for nameko services with consistent hashing support for multi-node setups. Uses the high-performance pymemcache library with automatic failover and connection pooling.

Key Features:

• Consistent hashing for reliable multi-node memcached clusters
• Optimized for Nameko - proper connection management and worker cleanup
• Django compatible - same interface as django-pymemcache for easy integration
• Automatic failover when nodes become unavailable
• Drop-in replacement for bmemcached with better multi-node behavior

Inspiration and structure proudly stolen from nameko-redis :) Thanks guys!

Installation

pip install nameko-pymemcache

Usage

from nameko.rpc import rpc
from nameko_pymemcache import Memcached


class MyService(object):
    name = "my_service"

    memcached = Memcached()

    @rpc
    def hello(self, name):
        self.memcached.set("foo", name)
        return "Hello, {}!".format(name)

    @rpc
    def bye(self):
        name = self.memcached.get("foo")
        return "Bye, {}!".format(name)

To specify memcached uri(s) and optional username/password you will need a config

AMQP_URI: 'amqp://guest:guest@localhost'
MEMCACHED_URIS: ['127.0.0.1:11211', ]
MEMCACHED_USER: 'playerone'
MEMCACHED_PASSWORD: 'ready'

Multi-Node Configuration

For multi-node memcached clusters, simply specify multiple servers:

AMQP_URI: 'amqp://guest:guest@localhost'
MEMCACHED_URIS: 
  - '192.168.1.10:11211'
  - '192.168.1.11:11211'
  - '192.168.1.12:11211'

The client automatically uses consistent hashing to distribute keys across nodes. When a node fails, only the keys on that node are affected (not all keys like with simple round-robin).

Advanced Configuration

You can pass extra options to customize client behavior:

class MyService(object):
    name = "my_service"

    memcached = Memcached(
        connect_timeout=0.1,    # connection timeout in seconds
        timeout=0.2,            # operation timeout in seconds
        retry_attempts=2,       # number of retries on failure
        dead_timeout=10,        # how long to avoid a failed node
    )

    ...

Framework Compatibility

This package provides a consistent interface that works seamlessly with Django's pymemcache backend, making it easy to share cache configurations between Django and Nameko services:

# Same server configuration works in both Django and Nameko
CACHE_SERVERS = ['192.168.1.10:11211', '192.168.1.11:11211', '192.168.1.12:11211']

# Django settings.py
CACHES = {
    'default': {
        'BACKEND': 'django.core.cache.backends.memcached.PyMemcacheCache',
        'LOCATION': CACHE_SERVERS,
    }
}

# Nameko config.yaml  
MEMCACHED_URIS: ['192.168.1.10:11211', '192.168.1.11:11211', '192.168.1.12:11211']

Performance Tips

• Identical server order: Keep the same server order across all clients for consistent key distribution
• Connection pooling: Available via pymemcache options if needed
• Custom timeouts: Override defaults by passing pymemcache options to the constructor
• Failure handling: Failed nodes are automatically removed from the hash ring and retried later