This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.

Bug 233262 - Big slowdown using DEV201307222300
Summary: Big slowdown using DEV201307222300
Status: RESOLVED FIXED
Alias: None
Product: editor
Classification: Unclassified
Component: Parsing & Indexing (show other bugs)
Version: 7.4
Hardware: PC Windows 7
: P2 normal (vote)
Assignee: Tomas Zezula
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-07-23 19:22 UTC by ecerichter
Modified: 2013-08-10 19:10 UTC (History)
1 user (show)

See Also:
Issue Type: DEFECT
Exception Reporter:


Attachments
Profiling snapshot (91.63 KB, application/octet-stream)
2013-07-23 19:22 UTC, ecerichter
Details
IDE log (559.92 KB, application/octet-stream)
2013-07-23 19:23 UTC, ecerichter
Details
New thread dump - machina had 90% of 8 core processor busy! (32.50 KB, application/octet-stream)
2013-07-31 01:25 UTC, ecerichter
Details
NB Self Sampler during very busy session (1.86 MB, application/zip)
2013-08-01 04:05 UTC, ecerichter
Details
Messages log... I don't know why it has been truncated? (39.40 KB, application/octet-stream)
2013-08-01 04:15 UTC, ecerichter
Details
New self samples: IDE doing nothing, with high CPU usage (381.59 KB, application/octet-stream)
2013-08-01 19:54 UTC, ecerichter
Details
IDE log (220.79 KB, application/octet-stream)
2013-08-01 19:56 UTC, ecerichter
Details
Prior IDE log (1.00 MB, application/octet-stream)
2013-08-01 19:56 UTC, ecerichter
Details
Sundelly, IDE cpu usage is low - here is the self sampler with low cpu usage (211.62 KB, application/octet-stream)
2013-08-01 19:58 UTC, ecerichter
Details

Note You need to log in before you can comment on or make changes to this bug.
Description ecerichter 2013-07-23 19:22:06 UTC
Created attachment 137641 [details]
Profiling snapshot

I've a big slowdown using DEV201307222300.
Win7 Pro, x64, JDK 1.7.0_25 x64.

Attached is IDE profiling snapshot and also IDE log.

Regards,

Edson
Comment 1 ecerichter 2013-07-23 19:23:19 UTC
Created attachment 137642 [details]
IDE log
Comment 2 Tomas Hurka 2013-07-24 08:04:23 UTC
Please describe what you did and what is slow and why do you think there is a problem. If you can provide steps to reproduce, it would be great. Thanks.
Comment 3 ecerichter 2013-07-24 13:27:03 UTC
I did nothing in special: just upgraded from DEV201307212300 to DEV201307222300 using ZIP archives.

Sundelly, my NetBeans started to raise processor from ~2% to 20%, and memory consumption raised from nearly 500Mb to 2Gb.
My current NB configuration says:

netbeans_default_options="--locale en:US -J-Xms1g -J-Xmx2g -J-client -J-XX:+UseParNewGC -J-XX:ParallelGCThreads=2 -J-XX:+PrintCommandLineFlags -J-XX:+UseConcMarkSweepGC -J-XX:+CMSClassUnloadingEnabled -J-XX:+ExplicitGCInvokesConcurrent -J-ea -J-Dapple.laf.useScreenMenuBar=true -J-Dapple.awt.graphics.UseQuartz=true -J-Dsun.zip.disableMemoryMapping=true -J-Dplugin.manager.check.updates=false -J-Dnetbeans.extbrowser.manual_chrome_plugin_install=yes"

I use mostly WebApps development (with or without JSF).

Nothing else changed: same JDK, same OS, no updates, no changes in anti-virus.
That's why I just collected a IDE profiling and IDE log.

I do have 5 servers registered: 3 Tomcat (all in different ports to avoid conflicts) and 2 Glassfish (3.1.2.2 and 4.0).

As you can see, there are lots of errors in IDE log.

I just can't tell what the problem is.
Comment 4 Tomas Hurka 2013-07-24 13:38:46 UTC
I see. It looks like most of the time is spent in output tab. Maybe one of the registered web servers prints a lot of log messages there. Can you, please, take a heap dump when memory consumption is at 2G, zip it  and upload it here <http://deadlock.netbeans.org/job/upload/build?delay=0sec>. Maybe this will show us , what is going on. Thanks.
Comment 5 ecerichter 2013-07-30 01:33:59 UTC
I've just got the slowdown again... this time, I've captured a complete set of heap dump, thread dump, and IDE log.
I've uploaded the heap dump in Hudson as you have asked.
Comment 6 Tomas Hurka 2013-07-30 07:30:42 UTC
(In reply to comment #5)
> I've just got the slowdown again... this time, I've captured a complete set of
> heap dump, thread dump, and IDE log.
> I've uploaded the heap dump in Hudson as you have asked.
I took a look at the heap dump and I do not see anything wrong. BTW: The heap dump is just 500M and not 2G you mentioned in previous message. Where is IDE log and thread dump?
Comment 7 ecerichter 2013-07-31 01:25:55 UTC
Created attachment 138023 [details]
New thread dump - machina had 90% of 8 core processor busy!
Comment 8 ecerichter 2013-07-31 02:05:20 UTC
I'm looking for the IDE log, but I'm afraid it was overwriten...
As soon as I get it, I'll attach it here.
Comment 9 Tomas Hurka 2013-07-31 13:28:29 UTC
(In reply to comment #7)
> Created attachment 138023 [details]
> New thread dump - machina had 90% of 8 core processor busy!
Thread dump is generally not very useful in such situation. Self sampling provides much better information. In general if you have high CPU usage, we would like to have .npss data from self sampler and messages.log. Make sure that both are from the same IDE run. If you see unusual heap consumption, we would like to inspect heap dump and messages.log.
Comment 10 ecerichter 2013-08-01 04:05:56 UTC
Created attachment 138097 [details]
NB Self Sampler during very busy session
Comment 11 ecerichter 2013-08-01 04:15:14 UTC
Created attachment 138098 [details]
Messages log... I don't know why it has been truncated?
Comment 12 Tomas Hurka 2013-08-01 06:45:04 UTC
(In reply to comment #11)
> Created attachment 138098 [details]
> Messages log... I don't know why it has been truncated?
The messages.log is truncated, because it is too long. The remaining part of the logs from your session will be probably in messages.log.1. Anyway this long log problem should be fixed now. See issue #233401.
Comment 13 Tomas Hurka 2013-08-01 06:49:12 UTC
(In reply to comment #10)
> Created attachment 138097 [details]
> NB Self Sampler during very busy session
npss snapshot shows that "Java Task Provider" thread is running all the time. Reassigning to "parsing & indexing". Please evaluate and reassign to appropriate component.
Comment 14 ecerichter 2013-08-01 19:54:53 UTC
Created attachment 138152 [details]
New self samples: IDE doing nothing, with high CPU usage

This is a npss I took when my IDE was doing virtually nothing (just open while I was drinking a cup of coffee).

CPU usage is high.
Comment 15 ecerichter 2013-08-01 19:56:13 UTC
Created attachment 138154 [details]
IDE log
Comment 16 ecerichter 2013-08-01 19:56:44 UTC
Created attachment 138155 [details]
Prior IDE log
Comment 17 ecerichter 2013-08-01 19:58:23 UTC
Created attachment 138156 [details]
Sundelly, IDE cpu usage is low - here is the self sampler with low cpu usage
Comment 18 Tomas Zezula 2013-08-02 10:06:17 UTC
Seems to be related to task list?
Do you have an Action Items window opened?
Comment 19 ecerichter 2013-08-02 12:17:57 UTC
(In reply to comment #18)
> Seems to be related to task list?
> Do you have an Action Items window opened?

Yes, I usually have the Action Items window open.
Comment 20 Tomas Zezula 2013-08-06 12:41:44 UTC
Fixed jet-main 0a9409ab0684

I was able to reduce number of source path finds ten times.
However I am not able to reproduce the problem. The error cache was not changed for ages and there was also no change in Action Items  according to the task list owner. I also added logging to find out the code causing the non needed error list refreshes.

Please wait until the fix is integrated, run the IDE with following cmd option and attache the log:

-J-Dorg.netbeans.modules.parsing.impl.indexing.errors.TaskProvider.level=FINER

It would be very helpful as we will able to identify the code causing the refreshes and possibly to
make it even faster.
Thanks!
Comment 21 Quality Engineering 2013-08-09 10:48:29 UTC
Integrated into 'main-silver', will be available in build *201308090746* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress)

Changeset: http://hg.netbeans.org/main-silver/rev/0a9409ab0684
User: Tomas Zezula <tzezula@netbeans.org>
Log: #233262:Big slowdown using DEV201307222300
Comment 22 ecerichter 2013-08-10 19:10:24 UTC
Verified. OK in DEV201308090746.
If needed, I'll reopen the ticket.
Thank you very much, this bug was tricky.