Redis is a caching method which can increase the speed of the backend and frontend of your shop. On Hypernode every customer has access to Redis cache, starting from 64 MB, depending on the plan. This article will explain how to configure Redis on your Magento 1 shop on Hypernode and how to work with redis-cli.

NB: When you used the hypernode-importer and you were already using Redis you don’t have to follow this tutorial.

Want to know how to configure Redis in Magento 2? Have a look at this article!

Download the Redis extension from Collin Mollenhour from Github

NB: If you’re using a Magento version later than 1.8.x, you can skip this extension. Magento 1.8 and bigger ship with a built-in Redis module included.

Follow these steps to change your caching backend to Redis:

  • Go to your public folder: /data/web/public
  • Use the following commands to install the Redis module on your shop:
    • modman init
    • modman clone git://github.com/colinmollenhour/Cm_Cache_Backend_Redis.git
  • Redis is now installed on your shop.

Change your local.xml file

NB: If you’re using a Magento version newer than 1.8.x you can use the built-in Redis module. To do so, replace <backend>Cm_Cache_Backend_Redis</backend> with <backend>Mage_Cache_Backend_Redis</backend> in the configuration snippet below. If you choose to use Cm_Cache_Backend_Redis when on 1.8.x, use the config snippet as is and install Cm_Cache_Backend_Redis.

Open your local.xml file and paste the next lines of code after the </crypt> tag:

Flush your cache after making these adjustments:

Use the redis-cli command

On Hypernode you can use the redis-cli command to get more information out of the Redis server. All the available commands can be found on the Command page of the Redis website. Below are some examples:

  • Flush all keys in all databases:

  • Flush all keys in database 0 or database 1 respectively:

  • Check how much Redis Cache you’re currently using

  • List all stored keys

Choosing the right database

If you use both a staging environment and a production site, pick the databases carefully. Using the same Redis databases for both production and development can cause unexpected behaviour on both sites.

Example:

Environment Db Type Redis Database
production Magento Cache 0
production Magento FPC 1
production Sessions 2
staging Magento Cache 3
staging Magento FPC 4
staging Sessions 5

Changing the compression library

It is possible to use the compression library ‘Snappy’ on Hypernode. More information about Snappy can be found in the changelog: Release-4224.

In order to use the compression library Snappy for your Redis cache you need to change <compression_lib>gzip</compression_lib> to <compression_lib>snappy</compression_lib>in your local.xml.

Configure Magento 1 to use Redis as the session store

You can use Redis for storing sessions too! To do so on Magento 1, read the article about storing sessions in Redis

Troubleshooting

A quick note, when you run into the configured max memory limit make sure that the necessary Redis keys are set to volatile (ensure an expire). Otherwise the complete allocated configured memory will fill up and Redis will ‘crash’.

When your Redis instance memory is full, and a new write comes in, Redis evicts keys to make room for the write based on your instance’s maxmemory policy. This is called the eviction policy.

In some cases we see that when Redis reaches the configured limit and tries to expire keys to make room, the eviction policy gets stuck in a loop. This means keys won’t be expired and Redis reaches it limit.

A temporary solution is to flush the Redis cache, you can do this by using the flushall command:

This will flush out all available Redis databases. Please keep in mind that this is only a temporary solution. The underlying cause is in the code of your application and needs to be permanently resolved.

Bots

As you know, the sessions of your webshop can also be stored in Redis. If you use Redis caching and store the sessions in Redis as well, you’ll have to share the available Redis memory. This shouldn’t be a problem on a regular basis, however we’ve seen scenarios wherein a shop stores its sessions in Redis and had some aggressive bots/crawlers visiting the shop. This resulted in a much larger amount of sessions to be stored in Redis than usual which is causing the Redis memory to fill up in no time, and crashes Redis.

You can check the bot traffic on your shop at any time on MageReport. If you want to get a more detailed insight in the bot traffic you can use the command pnl --yesterday --php --bots --fields ua | sort | uniq -c | sort -n to get an overview of the top 10 bots that visited your webshop yesterday. For more information about abuse bot check our article.

40