February 2011 Archives

What's New in WebGUI 8.0 #4 -- CHI Cache

Caching is a tricky business. Having just one kind of cache won't work, because the production environment will greatly determine the most efficient caching system. A distributed production environment would be best-served with a distributed cache. A smaller, single-server environment could use a simple shared memory cache.

Enter Jonathan Swartz's CHI module, the greatest Perl module to provide a unified caching interface. CHI is the DBI of caching: It presents an API, and delegates to CHI::Driver modules to perform the heavy lifting. It provides a layered caching system, allowing you to have a faster, more volatile cache in front of a slower, more persistent cache. It also provides a variable expiration time, preventing a "miss stampede" where all processes try to recompute an expired cache item at the same time.

By integrating CHI cache into WebGUI, we have the ability to provide any caching strategy that CHI can provide. We get Memcached, FastMmap, and DBI drivers (and more drivers can be written).

I wrote a CHI cache driver for WebGUI 7.9 that we've been using on many of our shared hosting servers. The performance increase using FastMmap through CHI over the old Storable+DBI cache module is dramatic: 2-5 times faster with CHI and FastMmap.

Using CHI in WebGUI

The fewer wrappers that WebGUI has around CPAN modules we use, the less code I have to write, and the more features will be available to our users without having to change WebGUI to use them.

To that end, you can write a section of the configuration file that gets passed directly to CHI->new. Some massaging occurs to make sure a DBI cache driver gets the right $dbh, but otherwise you can fully configure CHI directly from the WebGUI config file:

# The new default cache for WebGUI, FastMmap
{
     cache : {
         driver : 'FastMmap',
         root_dir : '/tmp/WebGUICache',
         expires_variance : 0.5
     }
 }

 # Set up a memcached cache with local memory in front
 {
     cache : {
         driver : 'Memcached::libmemcached',
         servers : [ '10.0.0.100:11211', '10.0.0.110:11211' ],
         l1_cache : {
            driver : 'Memory'
         }
     }
 }

When you want to use the cache in your code, you can get a CHI object with $session->cache. CHI's interface is sufficiently simple, with some fun tricks:

my $cache = $session->cache; # as read
my $value = $cache->get('cache_key');
if ( !$value ) {
    $value = compute_value();
    $cache->set( 'cache_key', $value );
}

# Combine get and set with intelligence
my $value = $cache->compute( 'cache_key', \&compute_value );

Future Plans

With a single unified cache that performs well and layers like CHI, we can take our current stow and scratch APIs and move them to the cache. In the case of stow, we remove a redundant API. In the case of scratch, we remove database hits.

We've also been exploring cache-only sessions, instead of updating the session every time a page is requested, updating the cache only, flushing to the database (or not). The fewer DB calls we make per page, the better performance will be.

Special thanks go out to Jonathan Swartz for such a wonderful solution.

Stay tuned for next time when I explore our new Admin Interface. Lots of pretty and screenshots!

What's New in WebGUI 8.0 #3: Upgrade System

Following The Path

If you installed WebGUI 0.9.0 back in August of 2001 (the first public release), you've had a stable upgrade path through WebGUI 7.10.8 (January 2011) and beyond. Plainblack.com has been through every upgrade for the last 10 years, a shining bastion to our upgradability.

A WebGUI 7.10 user would not even recognize a WebGUI 6.0 database, much less the database used by the 1.x series, but slowly, gradually, our upgrade system brought new features to every WebGUI site that wanted them.

The Ancient Way

Our old upgrade system was quite simple:

docs/upgrade_2.9.0-3.0.0.pl
docs/upgrade_3.0.0-3.0.1.sql
docs/upgrade_3.0.0-3.0.1.pl

Our upgrade.pl script would check for docs/upgrade_*, compare version numbers, and then execute the .sql and .pl scripts in order until there were no more upgrades left.

Because each .pl script was executed individually, there was a considerable amount of boilerplate in each script (123 lines). Because there was only one script per version, some scripts could get quite long. We had conventions to manage these limitations, but it was still a bit of a mind-twist to write an upgrade routine.

Later, when we moved to simultaneous beta and stable trees, it became even more difficult to manage these huge upgrade scripts. Collecting the new features from the beta tree to apply to the stable tree was a time-consuming manual task that some poor coder had to perform, back hunched over a dimly-lit screen in the wee hours of the night, testing and re-testing the upgrade to make sure stable lived up to its expectations.

Though our upgrade system had performed admirably, it was time for a fresh look at the problem.

The Modern Vision

The individual files for upgrades was working quite well, but didn't go far enough. Our new upgrade system has one file per upgrade step. Each sub from an old upgrade script would be one file in the new upgrade system. What's more, additional file types would be supported:

$ ls share/upgrades/7.10.4-8.0.0/
addNewAdminConsole.pl
admin_console.wgpkg
facebook_auth.sql
migrateToNewCache.pl
moveMaintenance.pl
moveRequiredProfileFields.pl

So now, instead of a single file for an upgrade, we have an entire directory. In this directory, the .pl files are scripts to be run, the .wgpkg files are WebGUI assets to add to the site, the .sql files are SQL commands to run, and any .txt files will be shown as a confirmation message to the user for gotchas like "All your users have been logged out as a result of this upgrade. Deal with it.".

So now, if you want to add your own custom upgrade routine, you just add another file to the directory which means less worrying about conflicts. When we need to build another new stable version release, we can just move the unique upgrade files from beta to the new upgrade.

The best part of the new upgrade system is how the .pl scripts are written. When you are in a .pl, you have a bunch of sugar to make the basic tasks much easier.

# Old upgrade routine. Just another day in a session
sub migrateToNewCache {
    my $session = shift;
    print "\tMigrating to new cache " unless $quiet;

    use File::Path;
    rmtree "../../lib/WebGUI/Cache";
    unlink "../../lib/WebGUI/Workflow/Activity/CleanDatabaseCache.pm";
    unlink "../../lib/WebGUI/Workflow/Activity/CleanFileCache.pm";

    my $config = $session->config;
    $config->set("cache", {
        driver              => 'FastMmap',
        expires_variance   => '0.10',
        root_dir            => '/tmp/WebGUICache',
    });

    $config->set("hotSessionFlushToDb", 600);
    $config->delete("disableCache");
    $config->delete("cacheType");
    $config->delete("fileCacheRoot");
    $config->deleteFromArray("workflowActivities/None", "WebGUI::Workflow::Activity::CleanDatabaseCache");
    $config->deleteFromArray("workflowActivities/None", "WebGUI::Workflow::Activity::CleanFileCache");

    my $db = $session->db;
    $db->write("drop table cache");
    $db->write("delete from WorkflowActivity where className in ('WebGUI::Workflow::Activity::CleanDatabaseCache','WebGUI::Workflow::Activity::CleanFileCache')");
    $db->write("delete from WorkflowActivityData where activityId in  ('pbwfactivity0000000002','pbwfactivity0000000022')");

    print "DONE!\n" unless $quiet;
}

If you're familiar with WebGUI session, this is pretty standard, but still much boilerplate and convention. The new scripts remove boilerplate and enforce what was once merely convention.

# New upgrade routine. migrateToNewCache.pl
use WebGUI::Upgrade::Script;
use Module::Find;

start_step "Migrating to new cache";

rm_lib
    findallmod('WebGUI::Cache'),
    'WebGUI::Workflow::Activity::CleanDatabaseCache',
    'WebGUI::Workflow::Activity::CleanFileCache',
;

config->set("cache", {
    'driver'            => 'FastMmap',
    'expires_variance'  => '0.10',
    'root_dir'          => '/tmp/WebGUICache',
});

config->set('hotSessionFlushToDb', 600);
config->delete('disableCache');
config->delete('cacheType');
config->delete('fileCacheRoot');
config->deleteFromArray('workflowActivities/None', 'WebGUI::Workflow::Activity::CleanDatabaseCache');
config->deleteFromArray('workflowActivities/None', 'WebGUI::Workflow::Activity::CleanFileCache');

sql 'DROP TABLE IF EXISTS cache';
sql 'DELETE FROM WorkflowActivity WHERE className in (?,?)',
    'WebGUI::Workflow::Activity::CleanDatabaseCache',
    'WebGUI::Workflow::Activity::CleanFileCache',
;
sql 'DELETE FROM WorkflowActivityData WHERE activityId IN (?,?)',
    'pbwfactivity0000000002',
    'pbwfactivity0000000022',
;

done;

The first thing we do in our new upgrade script is use WebGUI::Upgrade::Script. Now, instead of using the session for everything, we have subs imported for various tasks. This means that many times we can run an entire upgrade script without opening a WebGUI session, or creating a version tag unnecessarily.

If we do need a session, or a version tag, they will be automatically assigned relevant information describing what we're doing. When we're done, they will be automatically cleaned up and committed. What once was done with boilerplate, and subject to random deletion or subversion, is now enforced policy.

In all other respects, a WebGUI upgrade script is a Perl script. You can add modules, write subroutines, and do anything necessary to move WebGUI into the future.

The Internet is always evolving. With the WebGUI 8 upgrade system, we've made it easier to evolve with it.

Stay tuned for next time where I'll show off our CHI-based caching system.

About preaction

user-pic