|
|
Subject: Re: [opennms-devel] Crashing on Attempt to Connect - msg#00003
List: network.opennms.bugs
On Nov 3, 2004, at 2:53 PM, Judson Lester wrote:
That'd be great, except I don't have permission to upload files.
Username is "nyarly"
Strike that. File uploaded.
Judson
-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Was this page helpful?
Thread at a glance:
Previous Message by Date:
click to view message preview
Re: [opennms-devel] Crashing on Attempt to Connect
On Nov 2, 2004, at 8:09 AM, David Hustace wrote:
On Oct 29, 2004, at 5:47 PM, Judson Lester wrote:
On Oct 26, 2004, at 7:07 PM, Matt Brozowski wrote:
The bug that I saw was related to a bug in the JVM when garbage
collecting and the 'young generatation' was at 100%. It seemed to
be related to having many threads and connections being garbage
collected. We exercised it in 1.1.3 because 1.1.3 created an excess
of threads and would allow database connections to be garbage
collected without properly closing them.
These problem have been resolved in 1.1.4
Hrm. Well, I've done a complete clean install of 1.1.4. I still get
the very same JVM Internal Error about 'generation for allocation.'
The bug that Matt references only occurred in the OpenNMS service JVM;
not in the Tomcat JVM. This is the first that I have heard of this
error being reported in the Tomcat JVM. I googled for that error and
didn't find one link related to Java. Is that the exact error?
'generation for allocation'
Yes. The actual JVM error message occurs at the end of my log. It's
something like:
"Error occured in: generation for allocation" which is pretty cryptic
IMO. OTOH, none of the related JVM errors I've seen include a line
anything like that, so maybe some information is better than none.
Now, once in a while I can get it to log in successfully.
Once in a while certainly makes this issue much harder.
It's seldom enough that it's almost possible to ignore the successes.
It looks like some kind of race condition, where the race is almost
always lost. The times I have gotten in, it's been by explicitly
making changes to the java command line options. Adding -Xcheck:jni,
for instance, worked. Once.
Do want me to send logs directly to you, or to the list?
Upload the relevant logs to this file gallery on the OpenNMS Wiki so
everyone can get a look at them.
http://wiki.opennms.org/tiki-list_file_gallery.php?galleryId=4
That'd be great, except I don't have permission to upload files.
Username is "nyarly"
Judson
-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Next Message by Date:
click to view message preview
[opennms-devel] Proposed 1.2 Plan
Well... 1.1.4 has finally shipped... I think that Tarus has gone thru
and found all the patches that need to be applied based on the bugs
that are open. A preliminary list of the bugs and enhancements that we
believe should be fixed/developed before we ship 1.2 has been added as
dependencies of bug #804. This list includes an enhancement for
reformatting the code.
We could use some help testing 1.1.4 and CVS head as we finish these
changes for all of those that can help... I've marked with an asterisk
all of the items I could really use help with below.
The current rough development schedule (barring any serious problems)
is as follows: (You can add yourself as a CC to bug #804 to keep
abreast of progress)
This week:
get all the enhancements except for (possibly) #731 complete and
checked in
* continue testing 1.1.4/CVS to ensure that we didn't break it
Next week:
finish #731
fix remaining bugs
* test enhancements and continue regression testing
Week after:
Test as much as possible
create a stable branch in CVS for 1.2
build and release 1.2
* test enhancements and bug fixes.
Afterwards:
We will continue to address bug fixes in 1.2 but move away from larger
changes so 1.2 remains stable
Begin discussions of where OpenNMS needs to be in the future and start
the development of 1.3
Any feedback on this will be greatly appreciated... just thought I
would keep the development community aware of our plans...
Matt Brozowski
________________________________________________________________________
___
Matt Brozowski, OpenNMS Maintainer Main: +1 919 812 4984
The OpenNMS Group, Inc. Fax: +1 503
961 7746
Email: brozow@xxxxxxxxxxx URL:
http://www.opennms.com
-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Previous Message by Thread:
click to view message preview
Re: [opennms-devel] Crashing on Attempt to Connect
On Nov 2, 2004, at 8:09 AM, David Hustace wrote:
On Oct 29, 2004, at 5:47 PM, Judson Lester wrote:
On Oct 26, 2004, at 7:07 PM, Matt Brozowski wrote:
The bug that I saw was related to a bug in the JVM when garbage
collecting and the 'young generatation' was at 100%. It seemed to
be related to having many threads and connections being garbage
collected. We exercised it in 1.1.3 because 1.1.3 created an excess
of threads and would allow database connections to be garbage
collected without properly closing them.
These problem have been resolved in 1.1.4
Hrm. Well, I've done a complete clean install of 1.1.4. I still get
the very same JVM Internal Error about 'generation for allocation.'
The bug that Matt references only occurred in the OpenNMS service JVM;
not in the Tomcat JVM. This is the first that I have heard of this
error being reported in the Tomcat JVM. I googled for that error and
didn't find one link related to Java. Is that the exact error?
'generation for allocation'
Yes. The actual JVM error message occurs at the end of my log. It's
something like:
"Error occured in: generation for allocation" which is pretty cryptic
IMO. OTOH, none of the related JVM errors I've seen include a line
anything like that, so maybe some information is better than none.
Now, once in a while I can get it to log in successfully.
Once in a while certainly makes this issue much harder.
It's seldom enough that it's almost possible to ignore the successes.
It looks like some kind of race condition, where the race is almost
always lost. The times I have gotten in, it's been by explicitly
making changes to the java command line options. Adding -Xcheck:jni,
for instance, worked. Once.
Do want me to send logs directly to you, or to the list?
Upload the relevant logs to this file gallery on the OpenNMS Wiki so
everyone can get a look at them.
http://wiki.opennms.org/tiki-list_file_gallery.php?galleryId=4
That'd be great, except I don't have permission to upload files.
Username is "nyarly"
Judson
-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Next Message by Thread:
click to view message preview
[opennms-devel] Proposed 1.2 Plan
Well... 1.1.4 has finally shipped... I think that Tarus has gone thru
and found all the patches that need to be applied based on the bugs
that are open. A preliminary list of the bugs and enhancements that we
believe should be fixed/developed before we ship 1.2 has been added as
dependencies of bug #804. This list includes an enhancement for
reformatting the code.
We could use some help testing 1.1.4 and CVS head as we finish these
changes for all of those that can help... I've marked with an asterisk
all of the items I could really use help with below.
The current rough development schedule (barring any serious problems)
is as follows: (You can add yourself as a CC to bug #804 to keep
abreast of progress)
This week:
get all the enhancements except for (possibly) #731 complete and
checked in
* continue testing 1.1.4/CVS to ensure that we didn't break it
Next week:
finish #731
fix remaining bugs
* test enhancements and continue regression testing
Week after:
Test as much as possible
create a stable branch in CVS for 1.2
build and release 1.2
* test enhancements and bug fixes.
Afterwards:
We will continue to address bug fixes in 1.2 but move away from larger
changes so 1.2 remains stable
Begin discussions of where OpenNMS needs to be in the future and start
the development of 1.3
Any feedback on this will be greatly appreciated... just thought I
would keep the development community aware of our plans...
Matt Brozowski
________________________________________________________________________
___
Matt Brozowski, OpenNMS Maintainer Main: +1 919 812 4984
The OpenNMS Group, Inc. Fax: +1 503
961 7746
Email: brozow@xxxxxxxxxxx URL:
http://www.opennms.com
-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
|
|