deploying a war file is 4x to 10x slower than an ejb jar when admin port is unsecured about glassfish

[glassfishrobot]

• Issue Imported From: https://github.com/javaee/glassfish/issues/7
• Original Issue Raised By:@glassfishrobot
• Original Issue Assigned To: @yaminikb
• Original Issue Closed By:@glassfishrobot

[glassfishrobot]

@glassfishrobot Commented
@edburns said:
I'm new to glassfish, can you please direct me how to secure the http admin
port? The long lag for deployment of webapps is killing me, and this workaround
is very attractive.

[glassfishrobot]

@glassfishrobot Commented
@edburns said:
I found an option in the admin console that appeared to be what I needed. When
I checked it, the following mod was made to my domain.xml:

Unfortunately, when I tried to start the domain, I receive the following error
in my log file, and the domain does not start.

[#|2005-07-14T09:40:37.724-0400|SEVERE|TestAbbrevProductName9.0|javax.enterprise.system.core|_ThreadID=10;com.sun.enterprise.admin.server.core.JmxConnectorLifecycle@d6d53c;com.sun.appserv.server.ServerLifecycleException:
Internal: null ssl configuration;|Service
com.sun.enterprise.admin.server.core.JmxConnectorLifecycle@d6d53c cannot be
initialized! : com.sun.appserv.server.ServerLifecycleException: Internal: null
ssl configuration|#]

[#|2005-07-14T09:40:37.726-0400|SEVERE|TestAbbrevProductName9.0|javax.enterprise.system.core|_ThreadID=10;|CORE5071:
An error occured during initialization
com.sun.appserv.server.ServerLifecycleException: Internal: null ssl configuration
at
com.sun.enterprise.admin.server.core.JmxConnectorLifecycle.onInitialization(JmxConnectorLifecycle.java:102)
at
com.sun.enterprise.server.ApplicationServer.onInitialization(ApplicationServer.java:236)
at com.sun.enterprise.server.PEMain.run(PEMain.java:300)
at com.sun.enterprise.server.PEMain.main(PEMain.java:236)
Caused by: com.sun.appserv.server.ServerLifecycleException: Internal: null ssl
configuration
at
com.sun.enterprise.admin.server.core.JmxConnectorLifecycle.configureJmxConnectorServerDriver(JmxConnectorLifecycle.java:188)
at
com.sun.enterprise.admin.server.core.JmxConnectorLifecycle.onInitialization(JmxConnectorLifecycle.java:98)
... 3 more

| #] |

For one thing, this seems like a bug. I can file it if you like. For another,
I'm not sure if this was the proper setting to tweak to enable the security of
which you speak.

Reverting this setting back to false fixed the problem but now I'm stuck with
the ultra-slow deployment.

[glassfishrobot]

@glassfishrobot Commented
@edburns said:
P.S. I'm running on macosx tiger.

[glassfishrobot]

@glassfishrobot Commented
qouyang said:
This bug is currently under investigation. Please see below for the workaround:
Turn security on for PE.
In domain.xml, search for "4848" or "4849".
Set security-enabled="true".

[glassfishrobot]

@glassfishrobot Commented
qouyang said:
Reassign to JF, who is also the owner of Reassign to JF, who is also the owner of bug 6273329..

[glassfishrobot]

@glassfishrobot Commented
jfarcand said:
This problem is the result of This problem is the result of bug 6296095. Some war file requires more than 30
requests to be deployed, instead of 3 or 4.

Fixing that problem will improve the performance. Re-assigning the bug to Kedar
for further analysis.

Once bug 6296095 is fixed, please re-assign the bug to me so I can re-do my
analysis and see if there is still performance hotspot on winXp.. Some war file requires more than 30
requests to be deployed, instead of 3 or 4.

Fixing that problem will improve the performance. Re-assigning the bug to Kedar
for further analysis.

Once This problem is the result of bug 6296095. Some war file requires more than 30
requests to be deployed, instead of 3 or 4.

Fixing that problem will improve the performance. Re-assigning the bug to Kedar
for further analysis.

Once bug 6296095 is fixed, please re-assign the bug to me so I can re-do my
analysis and see if there is still performance hotspot on winXp. is fixed, please re-assign the bug to me so I can re-do my
analysis and see if there is still performance hotspot on winXp.

[glassfishrobot]

@glassfishrobot Commented
@vkraemer said:
doing a status whiteboard test. there is nothing to see here folks. Move along.

[glassfishrobot]

@glassfishrobot Commented
km105526 said:
I will verify that this is fixed because of JF's changes to grizzly.
Vince, is it possible for you to do so as well?

[glassfishrobot]

@glassfishrobot Commented
km105526 said:
Whereas the deployment times have significantly improved,
here is something that Vince has observed:
The "web-app deployment" time is more than the "ejb-app deployment time".

This just reminds me of deployment backend doing some checks to enforce the
uniqueness of web context roots across the domain. This particular task is
relevant only for web apps. Is this why web-app deployment is slower?

Qingqing, any idea?

[glassfishrobot]

@glassfishrobot Commented
km105526 said:
I will also compare this with 8.1

[glassfishrobot]

@glassfishrobot Commented
@vkraemer said:
here are my observations.

1. the deployment time has improved significantly for web apps.
2. the sample webapp simple still deploys slower than the ear file for the
simple sample for stateful session beans (which DOES NOT include a web app)

The delta is in the 2-3 second range, but the total time to deploy the ear (on
my system, YMMV, etc) was 3-5 seconds, SO 2-3 seconds is 50-100% slower.

[glassfishrobot]

@glassfishrobot Commented
@vkraemer said:
Note: these times were for b13

[glassfishrobot]

@glassfishrobot Commented
@honghzzhang said:
Bug 6296095 is fixed and the war deployment performance is much better since
then. Lower the priority of this bug to P3 and will mark it fixed once we meet
the performance exit criteria. is fixed and the war deployment performance is much better since
then. Lower the priority of this bug to P3 and will mark it fixed once we meet
the performance exit criteria.

[glassfishrobot]

@glassfishrobot Commented
@honghzzhang said:
lower to P3 according to my previous comments.

[glassfishrobot]

@glassfishrobot Commented
@honghzzhang said:
Mark the bug as fix since we now met Beta performance exit criteria

[glassfishrobot]

@glassfishrobot Commented
Was assigned to km105526

[glassfishrobot]

@glassfishrobot Commented
This issue was imported from java.net JIRA GLASSFISH-7

[glassfishrobot]

@glassfishrobot Commented
Reported by @vkraemer

[glassfishrobot]

@glassfishrobot Commented
Marked as fixed on Thursday, May 17th 2007, 4:08:27 am