Speed up OpenSAML initialization ?

[Brent Putman]

On 2/27/15 3:43 PM, Tom Zeller wrote:
>> I think Scott is right, it's much more likely to be in something like the
>> XML provider init.  But I still have never seen it take anywhere near that
>> long.
> Sorry, right.
>
> I added a log statement at the end of
> org.opensaml.core.config.InitializationService#initialize(), and
> that's why I reported the ~7s.

Ok, but easier to just use the existing DEBUG logging in the
InitializationService, that's what it's there for.

>
>
> 2015-02-27 14:37:49,432 - INFO
> [org.opensaml.core.config.InitializationService:48] - Initializing
> OpenSAML using the Java Services API
> ...
> 2015-02-27 14:37:55,583 - INFO
> [org.opensaml.core.config.InitializationService:63] - Done
> initializing OpenSAML using the Java Services API


I believe you but: I just ran a test on my real 3.0.0 test IdP VM in
our data center.  It's a relatively small VM in fact: 2 GB RAM; only 1
core, equivalent of Intel(R) Xeon(R) CPU X5570  @ 2.93GHz.  Running
RHEL 5.11.

The entire IdP took just a smidgen over 7 seconds to fully load.  The
OpenSAML init part of that was 1.7 seconds in total; the
AlgorithmRegistry part of that was 126ms.  The other 5+ seconds was all
the other IdP stuff: Spring, services, etc.

Not to sound flippant, but: how old is your workstation/laptop.  :-)



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shibboleth.net/pipermail/dev/attachments/20150227/19638ae9/attachment.html