You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by "Tom Mueller (JIRA)" <ji...@apache.org> on 2013/05/02 00:00:18 UTC
[jira] [Created] (FELIX-4042) improve Felix.init time by modifying
the way the cache is loaded
Tom Mueller created FELIX-4042:
----------------------------------
Summary: improve Felix.init time by modifying the way the cache is loaded
Key: FELIX-4042
URL: https://issues.apache.org/jira/browse/FELIX-4042
Project: Felix
Issue Type: Improvement
Components: Framework
Affects Versions: framework-4.2.1
Reporter: Tom Mueller
The GlassFish server has 292 bundles and this number continues to increase. When starting the server, the entire bundle cache is read in by Felix in the Felix.init method by calling Felix.reloadBundle in a simple for loop. Currently, in GlassFish this takes 440 ms on a MBP which is about 10% of the entire server startup time.
With the availability of multi-threaded CPUs and memory-cached disk I/O, it might be possible to improve Felix framework initialization time by making this process multi-threaded. Unfortunately, the reloadBundle method locks the global lock for almost its entire execution, so simply multi-threading the loop will not help. The reloadBundle logic will need to be refactored to make multi-threading effective in improving performance.
Another idea is to somehow flag certain bundles for reloading before other bundles and then allow "lazy" loading of this cache. When GlassFish starts, there are some bundles that are needed immediately while others are not accessed until later.
The goal for this improvement request is to have a constant initialization time for Felix that is independent of the number of bundles that are in the cache.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira