Questions

Questions and Answers

4.5.1: How does an event such as a crash of one of the memcached servers handled by the memcached client?

There is no automatic handling of this. If your client fails to get a response from a server then it should fall back to loading the data from the MySQL database.

The client APIs all provide the ability to add and remove memcached instances on the fly. If within your application you notice that memcached server is no longer responding, your can remove the server from the list of servers, and keys will automatically be redistributed to another memcached server in the list. If retaining the cache content on all your servers is important, make sure you use an API that supports a consistent hashing algorithm. For more information, see Section 4.2.4, “memcached Hashing/Distribution Types”.

4.5.2: What's a recommended hardware config for a memcached server? Linux or Windows?

memcached is only available on Unix/Linux, so using a Windows machine is not an option. Outside of this, memcached has a very low processing overhead. All that is required is spare physical RAM capacity. The point is not that you should necessarily deploy a dedicated memcached server. If you have web, application, or database servers that have spare RAM capacity, then use them with memcached.

If you want to build and deploy a dedicated memcached servers, then you use a relatively low-power CPU, lots of RAM and one or more Gigabit Ethernet interfaces.

4.5.3: memcached is fast - is there any overhead in not using persistent connections? If persistent is always recommended, what are the downsides (for example, locking up)?

If you don't use persistent connections when communicating with memcached then there will be a small increase in the latency of opening the connection each time. The effect is comparable to use nonpersistent connections with MySQL.

In general, the chance of locking or other issues with persistent connections is minimal, because there is very little locking within memcached. If there is a problem then eventually your request will timeout and return no result so your application will need to load from MySQL again.

4.5.4: How expensive is it to establish a memcache connection? Should those connections be pooled?

Opening the connection is relatively inexpensive, because there is no security, authentication or other handshake taking place before you can start sending requests and getting results. Most APIs support a persistent connection to a memcached instance to reduce the latency. Connection pooling would depend on the API you are using, but if you are communicating directly over TCP/IP, then connection pooling would provide some small performance benefit.

4.5.5: How will the data will be handled when the memcached server is down?

The behavior is entirely application dependent. Most applications will fall back to loading the data from the database (just as if they were updating the memcached) information. If you are using multiple memcached servers, you may also want to remove a server from the list to prevent the missing server affecting performance. This is because the client will still attempt to communicate the memcached that corresponds to the key you are trying to load.

4.5.6: What is the max size of an object you can store in memcache and is that configurable?

The default maximum object size is 1MB. If you want to increase this size, you have to re-compile memcached. You can modify the value of the POWER_BLOCK within the slabs.c file within the source.

In memcached 1.4.2 and higher you can configure the maximum supported object size by using the -I command-line option. For example, to increase the maximum object size to 5MB:

$ memcached -I 5m

4.5.7: Can memcached be run on a Windows environment?

No. Currently memcached is available only on the Unix/Linux platform. There is an unofficial port available, see http://www.codeplex.com/memcachedproviders.

4.5.8: What are best practices for testing an implementation, to ensure that it is an improvement over the MySQL query cache, and to measure the impact of memcached configuration changes? And would you recommend keeping the configuration very simple to start?

The best way to test the performance is to start up a memcached instance. First, modify your application so that it stores the data just before the data is about to be used or displayed into memcached.Since the APIs handle the serialization of the data, it should just be a one line modification to your code. Then, modify the start of the process that would normally load that information from MySQL with the code that requests the data from memcached. If the data cannot be loaded from memcached, default to the MySQL process.

All of the changes required will probably amount to just a few lines of code. To get the best benefit, make sure you cache entire objects (for example, all the components of a web page, blog post, discussion thread, etc.), rather than using memcached as a simple cache of individuals rows of MySQL tables. You should see performance benefits almost immediately.

Keeping the configuration very simple at the start, or even over the long term, is very easy with memcached. Once you have the basic structure up and running, the only addition you may want to make is to add more servers into the list of servers used by your clients. You don't need to manage the memcached servers, and there is no complex configuration, just add more servers to the list and let the client API and the memcached servers make the decisions.

4.5.9: Can MySQL actually trigger/store the changed data to memcached?

Yes. You can use the MySQL UDFs for memcached and either write statements that directly set the values in the memcached server, or use triggers or stored procedures to do it for you. For more information, see Section 4.3.7, “Using the MySQL memcached UDFs”

4.5.10: So the responsibility lies with the application to populate and get records from the database as opposed to being a transparent cache layer for the db?

Yes. You load the data from the database and write it into the cache provided by memcached. Using memcached as a simple database row cache, however, is probably inefficient. The best way to use memcached is to load all of the information from the database relating to a particular object, and then cache the entire object. For example, in a blogging environment, you might load the blog, associated comments, categories and so on, and then cache all of the information relating to that blog post. The reading of the data from the database will require multiple SQL statements and probably multiple rows of data to complete, which is time consuming. Loading the entire blog post and the associated information from memcached is just one operation and doesn't involve using the disk or parsing the SQL statement.

4.5.11: Is compression available?

Yes. Most of the client APIs support some sort of compression, and some even allow you to specify the threshold at which a value is deemed appropriate for compression during storage.

4.5.12: File socket support for memcached from the localhost use to the local memcached server?

You can use the -s option to memcached to specify the location of a file socket. This automatically disables network support.

4.5.13: Are there any, or are there any plans to introduce, a framework to hide the interaction of memcached from the application; that is, within hibernate?

There are lots of projects working with memcached. There is a Google Code implementation of Hibernate and memcached working together. See http://code.google.com/p/hibernate-memcached/.

4.5.14: What are the advantages of using UDFs when the get/sets are manageable from within the client code rather than the db?

Sometimes you want to be able to be able to update the information within memcached based on a generic database activity, rather than relying on your client code. For example, you may want to update status or counter information in memcached through the use of a trigger or stored procedure. For some situations and applications the existing use of a stored procedure for some operations means that updating the value in memcached from the database is easier than separately loading and communicating that data to the client just so the client can talk to memcached.

In other situations, when you are using a number of different clients and different APIs, you don't want to have to write (and maintain) the code required to update memcached in all the environments. Instead, you do this from within the database and the client never gets involved.

4.5.15: Is memcached typically a better solution for improving speed than MySQL Cluster and\or MySQL Proxy?

Both MySQL Cluster and MySQL Proxy still require access to the underlying database to retrieve the information. This implies both a parsing overhead for the statement and, often, disk based access to retrieve the data you have selected.

The advantage of memcached is that you can store entire objects or groups of information that may require multiple SQL statements to obtain. Restoring the result of 20 SQL statements formatted into a structure that your application can use directly without requiring any additional processing is always going to be faster than building that structure by loading the rows from a database.

4.5.16: What speed trade offs is there between memcached vs MySQL Query Cache? Where you check memcached, and get data from MySQL and put it in memcached or just make a query and results are put into MySQL Query Cache.

In general, the time difference between getting data from the MySQL Query Cache and getting the exact same data from memcached is very small.

However, the benefit of memcached is that you can store any information, including the formatted and processed results of many queries into a single memcached key. Even if all the queries that you executed could be retrieved from the Query Cache without having to go to disk, you would still be running multiple queries (with network and other overhead) compared to just one for the memcached equivalent. If your application uses objects, or does any kind of processing on the information, with memcached you can store the post-processed version, so the data you load is immediately available to be used. With data loaded from the Query Cache, you would still have to do that processing.

In addition to these considerations, keep in mind that keeping data in the MySQL Query Cache is difficult as you have no control over the queries that are stored. This means that a slightly unusual query can temporarily clear a frequently used (and normally cached) query, reducing the effectiveness of your Query Cache. With memcached you can specify which objects are stored, when they are stored, and when they should be deleted giving you much more control over the information stored in the cache.

4.5.17: Does the -L flag automatically sense how much memory is being used by other memcached?

No. There is no communication or sharing of information between memcached instances.

4.5.18: Is the data inside of memcached secure?

No, there is no security required to access or update the information within a memcached instance, which means that anybody with access to the machine has the ability to read, view and potentially update the information. If you want to keep the data secure, you can encrypt and decrypt the information before storing it. If you want to restrict the users capable of connecting to the server, your only choice is to either disable network access, or use IPTables or similar to restrict access to the memcached ports to a select set of hosts.

4.5.19: Can we implement different types of memcached as different nodes in the same server - so can there be deterministic and non deterministic in the same server?

Yes. You can run multiple instances of memcached on a single server, and in your client configuration you choose the list of servers you want to use.

4.5.20: How easy is it to introduce memcached to an existing enterprise application instead of inclusion at project design?

In general, it is very easy. In many languages and environments the changes to the application will be just a few lines, first to attempt to read from the cache when loading data and then fall back to the old method, and to update the cache with information once the data has been read.

memcached is designed to be deployed very easily, and you shouldn't require significant architectural changes to your application to use memcached.

4.5.21: Can memcached work with ASPX?

There are ports and interfaces for many languages and environments. ASPX relies on an underlying language such as C# or VisualBasic, and if you are using ASP.NET then there is a C# memcached library. For more information, see .

4.5.22: If I have an object larger then a MB, do I have to manually split it or can I configure memcached to handle larger objects?

You would have to manually split it. memcached is very simple, you give it a key and some data, it tries to cache it in RAM. If you try to store more than the default maximum size, the value is just truncated for speed reasons.

4.5.23: How does memcached compare to nCache?

The main benefit of memcached is that is very easy to deploy and works with a wide range of languages and environments, including .NET, Java, Perl, Python, PHP, even MySQL. memcached is also very lightweight in terms of systems and requirements, and you can easily add as many or as few memcached servers as you need without changing the individual configuration. memcached does require additional modifications to the application to take advantage of functionality such as multiple memcached servers.

4.5.24: Doing a direct telnet to the memcached port, is that just for that one machine, or does it magically apply across all nodes?

Just one. There is no communication between different instances of memcached, even if each instance is running on the same machine.

4.5.25: Is memcached more effective for video and audio as opposed to textual read/writes

memcached doesn't care what information you are storing. To memcached, any value you store is just a stream of data. Remember, though, that the maximum size of an object you can store in memcached without modifying the source code is 1MB, so it's usability with audio and video content is probably significantly reduced. Also remember that memcached is a solution for caching information for reading. It shouldn't be used for writes, except when updating the information in the cache.

4.5.26: We are caching XML by serialising using saveXML(), because PHP cannot serialise DOM objects; Some of the XML is variable and is modified per-request. Do you recommend caching then using XPath, or is it better to rebuild the DOM from separate node-groups?

You would need to test your application using the different methods to determine this information. You may find that the default serialization within PHP may allow you to store DOM objects directly into the cache.

4.5.27: Do the memcache UDFs work under 5.1?

Yes.

4.5.28: Is it true memcached will be much more effective with db-read-intensive applications than with db-write-intensive applications?

Yes. memcached plays no role in database writes, it is a method of caching data already read from the database in RAM.

4.5.29: How are auto-increment columns in the MySQL database coordinated across multiple instances of memcached?

They aren't. There is no relationship between MySQL and memcached unless your application (or, if you are using the MySQL UDFs for memcached, your database definition) creates one.

If you are storing information based on an auto-increment key into multiple instances of memcached then the information will only be stored on one of the memcached instances anyway. The client uses the key value to determine which memcached instance to store the information, it doesn't store the same information across all the instances, as that would be a waste of cache memory.

4.5.30: If you log a complex class (with methods that do calculation etc) will the get from Memcache re-create the class on the way out?

In general, yes. If the serialization method within the API/language that you are using supports it, then methods and other information will be stored and retrieved.