Server goes into constant Full GC loop after running for a day

This is a servlet application running under Resin, with Java hotspot 1.5_13 on Linux 64bit OS. JVM args are:
-verbosegc -Xms8G -Xmx8G -XX:NewSize=512m -XX:MaxNewSize=512m -XX:+TraceClassUnloading -XX:+UseParallelGC -XX:PermSize=512M -XX:MaxPermSize=512M -XX:+PrintGCTimeStamps -XX:+PrintGCDetails -XX:+DisableExplicitGC
After running for a day or so, the server goes in to constant Full GC every few seconds, essentially hanging the server. Here is an excerpt of the JVM log:
127731.579: [Full GC [PSYoungGen: 156635K->0K(352576K)] [PSOldGen: 7751937K->2260734K(7864320K)] 7908573K->2260734K(8216896K) [PSPermGen: 62855K->62855K(524288K)], 5.2972428 secs]
127737.596: [GC [PSYoungGen: 179328K->157147K(351040K)] 2440062K->2417881K(8215360K), 0.1275439 secs]
127739.066: [GC [PSYoungGen: 336475K->155824K(355136K)] 2709605K->2683786K(8219456K), 0.1729460 secs]
127740.502: [GC [PSYoungGen: 340464K->169143K(353792K)] 2980821K->2836693K(8218112K), 0.1914048 secs]
127741.562: [GC [PSYoungGen: 353783K->166305K(345344K)] 3021333K->2860977K(8209664K), 0.1367291 secs]
127742.973: [GC [PSYoungGen: 340908K->153430K(349568K)] 3147976K->3124907K(8213888K), 0.1970560 secs]
127743.946: [GC [PSYoungGen: 328278K->156065K(349568K)] 3299755K->3277928K(8213888K), 0.1868634 secs]
127745.498: [GC [PSYoungGen: 330183K->156854K(349568K)] 3564442K->3545105K(8213888K), 0.1836370 secs]
From the log it seemed that the memory usage jumped by 150M to 250M within 1.5 second, even when the user traffic was very light (a couple of page views a second, vs. 10 page views a second at peak). The same app was running fine for many hours on a Windows dev machine in load testing condition with about 300M heapsize. No memory leakage was found so far. (it would be pretty unlikely to have such a large memory leakage without being noticed in dev and testing)
Thanks for any help.
Edited by: tuxzilla on Nov 16, 2008 11:43 PM
Edited by: tuxzilla on Nov 16, 2008 11:46 PM

A NewSize of 512m seems small for an 8G heap. This would probably cause most objects to be rapidly tenured into the OldGen heap space. (Ultimately depends on your application.) You might consider either using a NewRatio in place of NewSize/MaxNewSize or not use either and instead see how the jvm handles things itself with"-XX:+UseAdaptiveSizePolicy"
VisualVM is a good graphical tool for watching gc activity.
NewRatio=1 is the largest you can go for Eden, which will set (Eden and survivor spaces) equal to oldGen space. ie, for 8G, 4g each.
Watch activity in the gc tool, you will see "sawtooth" activity. If your sawtooth pattern does not maintain a flat profile, then you have in fact have a memory leak.

Similar Messages

Not permission to install Lion into my iMac HD even after running Disk Tools repairing permissions ???

Not permission to install Lion into my iMac HD even after running Disk Tools repairing permissions ???

Post the exact text of the error message you get.

HP B210a goes into a sleep mode if not used for several hours,

Running my B210a eprinter in windows vista. The printer goes into a sleep mode if not used for several hours. The power button is slowly flashing dim then bright, and I need to manually touch the printer's screen to enable it to print. If printer not used for several days, the only way I can make it work is to unplug and plug. Any suggestions or help appreciated. J Rice

What OS? Wired or wireless connection?
Say thanks by clicking the Kudos Thumbs Up to the right in the post.
If my post resolved your problem, please mark it as an Accepted Solution ...
I worked for HP but now I'm retired!

HELP - ABAP Program Cancelled After Running for 2.5 Hours in Background

I have an ABAP report program that was being tested in our QA system and it died after running for 2.5 hours with the status of "Cancelled" (which, I assume, means it was terminated by SAP for exceeding some type of governor).
I ran Code Analyzer and no performance issues were recognized.
I am running several SELECT statements during program execution. These are a list of all the SELECT statements used in my program. Also, I'm using PNPCE to get a list of pernrs for which to pull data at the beginning of the program.
SELECT *
 INTO CORRESPONDING FIELDS OF TABLE gt_p0167
 FROM pa0167
 WHERE pernr EQ gt_selected_pernrs
 AND bplan IN s_bplan
 AND begda LE pn-endda
 AND endda GE pn-begda.
 SELECT SINGLE fgbdt fasex
 INTO (lv_dob, lv_gender)
 FROM pa0021
 WHERE pernr EQ gt_selected_pernrs
 AND subty EQ <fs_dtyxx>
 AND objps EQ <fs_didxx>
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE agency ansvh
 INTO (p_agency_out, lv_ansvh)
 FROM pa0001
 WHERE pernr EQ p_pernr_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE vorna nachn gbdat perid gesch
 INTO (gt_control_table-fname, gt_control_table-lname, gt_control_table-dob, gt_control_table-mskssn, gt_control_table-gender)
 FROM pa0002
 WHERE pernr EQ p_pernr_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE perid
 INTO (lv_ssn)
 FROM pa0106
 WHERE pernr EQ p_pernr_in
 AND subty EQ p_subty_in
 AND objps EQ p_objps_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE smoke
 INTO (p_smoker_out)
 FROM pa0376
 WHERE pernr = p_pernr_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE state pstlz zcounty
 INTO (p_state_out, lv_zip, p_county_out)
 FROM pa0006
 WHERE pernr EQ p_pernr_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
SELECT SINGLE werks btrtl
 INTO (lv_werks, lv_btrtl)
 FROM pa0001
 WHERE pernr EQ p_pernr_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
 SELECT SINGLE region
 INTO (p_region_out)
 FROM zpat_county_code
 WHERE county EQ p_county_in.
 SELECT SINGLE pernr
 INTO lv_pernr
 FROM pa0002
 WHERE perid EQ gt_cobra_table-l_essn.
SELECT SINGLE eecst ercst
 INTO (lv_eecst, lv_ercst)
 FROM t5ubi
 WHERE barea EQ '01'
 AND bplan EQ lv_bplan
 AND bcost EQ p_bcost_in
 AND cstv1 EQ '0001'
 AND smoke EQ p_smoker_in
 AND begda LE pn-endda
 AND endda GE pn-begda.
 SELECT SINGLE kwert
 INTO lv_kwert
 FROM t511k
 WHERE molga EQ '10'
 AND konst EQ 'ZCOB1'
 AND begda LE pn-endda
 AND endda GE pn-begda.

The screen shot the tester sent me only says that it was canceled. However, the log shows the following:
09/17/2008 16:54:32 Job Started
09/17/2008 16:54:32 Step 001 started (program ZBNI001...
09/17/2008 19:22:09 ABAP/4 processor: DATASET_NOT_OPEN
09/17/2008 19:22:09 Job cancelled
I'm not sure it was an error, per se, but maybe SAP canceled it because it had been running too long? The "DATASET_NOT_OPEN" is curious although I've no idea what it means.
Thanks for your help.

HT3964 My computer freezes after running for about 15 min even after setting the energy saving to 3 hours

My Mac computer freezes after running for 15 min even after seeting the energy saving to 3 hours. Any ideas how to stop the freezing/

Hi Teodorafrombc,
Thanks for visiting Apple Support Communities.
If your computer is becoming unresponsive, you may want to see if specific applications are causing this behavior. Try using the Force Quit menu to check for unresponsive applications:
Using the Force Quit feature of OS X makes an application close, even when it is not responsive.
Important: Normally you should not need to force quit an app to close it. When an app is forced to quit, any unsaved changes to open documents are not saved, so try these methods to normally close the app first:
Choose Quit from the app menu. For example, in Safari, choose Safari > Quit Safari.
Choose Quit by right-clicking or control-clicking on an app's icon in the Dock
You can find the article with these steps here:
OS X: How to quit an unresponsive application using Force Quit
http://support.apple.com/kb/ht3411
If your display is going blank or seems to be going to sleep unexpectedly, try these steps first:
Reset the system
You can reset the Mac's parameter RAM and SMC.
You can find the article with these steps and more information here:
Apple computers: Troubleshooting issues with video on internal or external displays
http://support.apple.com/kb/ht1573
If these steps don't resolve the issue, feel free to reply with more information about the symptoms you are experiencing.
Best,
Jeremy

Mail quits after running for a few minutes

Mail quits after running for a couple of minutes, started doing it yesterday. I have looked at the advice given for similar problems, disk utility can find no problem with my drive, I repaired permissions anyway. No change.
Also ran the 'plutil' on the plist which came out as 'OK'.
I don't 'think' it is immediately the result of updating to 10.4.7 as it ran happily for some time after the update.
Is the only solution to trash the plist and redo all my mail accounts?
20' iMac 2.1 Ghz PowerPC G5 Mac OS X (10.4.7) 1 GB RAM

Mail probably is crashing when new junk mail arrives. Try the following in order to identify the problem:
1. Either tell Mail to go offline immediately after opening it, or shut down your Internet connection and open Mail afterwards.
2. Disable junk mail filtering in Mail > Preferences > Junk Mail.
3. Go online again.
If Mail doesn't crash now, then you know for sure there is something wrong with the junk filter. In that case, the following procedure should fix the problem:
1. Quit Mail if it's running.
2. In the Finder, go to ~/Library/Mail/.
3. Locate LSMMap2 and move it to the Trash -- deleting this file is what clicking the Reset button in Preferences > Junk Mail does. If you see an LSMMap file there, move it to the Trash too.
4. Locate DefaultCounts and move it to the Trash. This file keeps statistics about the number of messages that have been marked as junk, either automatically by the junk filter or manually by the user.
5. Locate MessageRules.plist and move it to the Desktop. This is where Mail 2.x stores all rules, including those used by the junk filter. If there is a file called MessageRules.plist.backup, move it to the Desktop too. You may also see MessageSorting.plist files there; this is where Mail 1.x stored the rules, and they are no longer used by Mail 2.x, so just move them to the Trash if you see them.
6. Look for the account folders. The name of each account folder starts with the account type (POP, IMAP, Mac), followed by the account username and the incoming mail server. Open each of the account folders, locate the Junk.mbox or Junk.imapmbox folder within it, and move it to the Trash if present.
7. Go to ~/Library/Mail/Mailboxes/. Locate any Junk.mbox folders there (their name may include the account name in parenthesis), and move them to the Trash.
Note: Messages in Junk folders with an .mbox suffix are stored locally, and will be lost forever as a result of doing steps 6 and 7. If there is a chance that you have legit messages there and you want to preserve them, move them to another mailbox in Mail first (you may prevent Mail from crashing by shutting down the Internet connection or telling Mail to go offline immediately after opening it). Messages in Junk folders with an .imapmbox suffix are stored on the server and can be ignored (Mail will rebuild those mailboxes automatically).
8. Open Mail. As a result of removing the rules file, the junk filter will be disabled now. You may want to either tell Mail to go offline immediately after opening it, or shut down the Internet connection before opening Mail, to prevent it from downloading anything until the junk mail filter has been enabled again.
9. Go to Mail > Preferences > Junk Mail, enable junk filtering, and configure it however you wish.
10. Go online again if you went offline in step 8.
If the above solves the problem and you don't have any rules you'd like to preserve or recreating them is not a daunting task, just delete the MessageRules.plist files that were moved to the Desktop and be done with it.
Note: For those not familiarized with the ~/ notation, it refers to the user's home folder. You can easily locate any of the folders referred to in this post by copying the file path here, doing Go > Go to Folder in the Finder, and pasting the file path there.

Firefox 3.6.13 on Mac 10.6 I lose the "mouseover" capability after running for a while. Only a REBOOT of the MAC will restore.

After running for a while (hours ? days?) I lose the ability to "mouseover" data on a website and get info. Specifically on Yahoo's
portal page, I can't mouse over a news item and get a summary.
Restarting Firefox, in safe mode or normal mode, does not restore the function. Rebooting the MAC (unheard of !!) restores. What gives ?

See if this works.
Type '''about:config''' in the URL bar and hit enter.
Filter = '''print.print''' 
There are 6 prefs for headers / footers. 
'''print.print_header...''' center, left, & right 
and 
'''print.print_footer...''' center, left, & right 
You can add what you want in the '''''Value''''' column and then Restart Firefox to effect the change.

IDM server goes into recovered state

Hi everybody,
When I install a second IDM server (on another machine) that shares the same Oracle repository as the first one, I notice that the first server goes from the "active" state to the "recovered" state. This is seen under the Configuration->Servers tab.
Any ideas why this happens?
Thanks!

When you are looking into? Can you provide more info? What does the process do?
Best Regards,Uri Dimant SQL Server MVP,
http://sqlblog.com/blogs/uri_dimant/
MS SQL optimization: MS SQL Development and Optimization
MS SQL Consulting:
Large scale of database and data cleansing
Remote DBA Services:
Improves MS SQL Database Performance
SQL Server Integration Services:
Business Intelligence

My (2008) macbook pro, running 10.5.8 runs for about 30 seconds and then goes into deep sleep mode. After significant difficulty awakening, it runs for 30 seconds and....start repeating here. It did this once before about a year ago.

.....and then goes into deep sleep mode. When it (finally) awakens....it runs for about 30 seconds and.....start repeating here.
And yes, I've shut dwn and restarted multiple times, rebooted in terminal mode and run fsck, rebooted and reset the parameter ram, yadda-yadda......
Anybdy got any ideas?

Power or other problems related to unexpected system sleep, shutdown, lights or fans call for an SMC reset. Read all the steps.
Before Resetting the SMC
Try each of the following steps in this order before you reset the SMC. Test the issue after completing each troubleshooting step to determine if the issue still occurs.
Press Command + Option + Escape to force quit any application that is not responding.
Put your Mac to sleep by choosing the Apple () menu from the upper-left menu bar and then choosing Sleep. Wake the computer after it has gone to sleep.
Restart your Mac by choosing the Apple () menu from the upper-left menu bar and then choosing Restart.
Shut down your Mac by choosing the Apple () menu from the upper-left menu bar and then choosing Shut Down.
Resetting the SMC on Mac portables with a battery you can remove
Shut down the computer.
Disconnect the MagSafe power adapter from the computer, if it's connected.
Remove the battery (to remove the battery - click here: MacBook or MacBook Pro).
Press and hold the power button for 5 seconds.
Release the power button.
Reconnect the battery and MagSafe power adapter.
Press the power button to turn on the computer.
Still having problems? Replace steps 6 and 7 with the following:
Reconnect just the MagSafe power adapter.
Press the power button to turn on the computer,
Then reconnect the battery.
Resetting the SMC on portables with a battery you should not remove on your own
Shut down the computer.
Plug in the MagSafe power adapter to a power source, connecting it to the Mac if it's not already connected.
On the built-in keyboard, press the (left side) Shift-Control-Option keys and the power button at the same time. The power adapter indicator light may cycle off / on once.
Release all the keys and the power button at the same time.
Press the power button to turn on the computer.

HELP!!!!! Z10 starts up but goes into a load up loop

When you say flash it what do you mean????

Hi Everyone, I NEED HELP!!!!! my Z10 didn't load my text messages this morning so i restarted it and now it goes into the loading screen gets to 100% and freezes every now and then restarts its self only to do the same again, i plugged it into my laptop and now that won't recognise it!!! any ideas on how to fix it??????

My phone is frozen....this is the second phone that has done this. How can I get my phone working without going into a store spending countless hours only for them to minimize the problem.

My phone froze this morning while I was replying to an email. Is there anything I can do to get it working and to avoid going into Apple?

Assuming it is charged and did not shutdown due to low battery, have you tried an power off/on? Or a reset by holding the power and home buttons until the apple logo appears and it restarts?

Why App World again blocked in Pakistan after working for 1 day

I'm user of Ufone-Pakistan Network, i got a message from my provider, about app world being enabled in the country and it also worked for 1 day, but again after that its blocked.
Any one knows when will app world services resume in Pakistan.

Hi Deepak,
The Archive server is running as usual. But the ECC was a bit slow than usual. No error msg. I solve the problem by forwarding the message through T-Code SWIA and the workflow is working fine. It just that i need to complete the workflow manually in T-Code SWIA after posting. How can i automatically complete the workflow?

Vi locks up after running for days

I have some old code written in LV 7 that I ported over to 2011. After fixing the legacy issues with DAQ and a save routine I ran it and it worked fine. Three days ago it was started in order to record TC data and this morning is froze/crashed. The other code which I did the same to runs for months with no incident. I checked for files becoming large with nothing out of the ordenary. Before re-launching LV and the vi I started WinDbg to record a dump file. Anyone have any ideas or ran into a simular problem.

Hello teslac,
Did you also update your OS or PC when migrating to LabVIEW 2011? Nothing specific stands out based on your comments, but those changes would also be worth evaluating. WinDbg is certainly a good idea. Additionally:
-Are you using LabVIEW 2011 or 2011 SP1? 32- or 64-bit?
-What OS is this? What bitness?
-Is this running in the LabVIEW development environment or is it compiled into an executable running on the machine?
-Did the application crash and close, or just become unresponsive?
-If the application did crash, a log file should have been generated by the IDE or runtime engine. If you have that available, I'd recommend posting that here as it may give some insight to the issue.
Regards,
Tom L.

My iPod touch 4g will not turn on after charging for a day what do I do ?

please help

Try:
- iOS: Not responding or does not turn on
- Also try DFU mode after try recovery mode
How to put iPod touch / iPhone into DFU mode « Karthik's scribblings
- If not successful and you can't fully turn the iOS device fully off, let the battery fully drain. After charging for an least an hour try the above again.
- Try another cable
- Try on another computer
- If still not successful that usually indicates a hardware problem and an appointment at the Genius Bar of an Apple store is in order.
Apple Retail Store - Genius Bar

Query on a transaction table breaks after running for sometime - Oracle 11g

Hi,
I'm running the below query to pull the sum of amount on trans table (which has 12 years of data partitioned yearly) whose last transaction date is 19-MAY-2013. The calendardatekey column is the one which is used in place of transaction date (each day has a unique calendardatekey). The query runs for couple of hours and breaks. When I contacted DBA they told that its trying to do a full table scan and pulling huge amount of data and doing a kind of sort and then breaks. Could someone please help if there is a better approach? If it works out for one day I will have to perform this for few more days.
Oracle version - 11g
trans table :
columns customerkey, programid and amount - are indexed.
card table :
column customerkey, status are indexed.
select t.customerkey, t.programid, sum(t.amount), max(calendardatekey) from trans t
where
exists (select null from card c where t.customerkey = c.customerkey and c.status = 'A')
--and t.programid = 10
group by t.customerkey, t.instprogkey
having sum(t.amount) < -100
and max(calendardatekey) =4888
Thanks for your help!

doubtsinora wrote:
Hi,
I'm running the below query to pull the sum of amount on trans table (which has 12 years of data partitioned yearly) whose last transaction date is 19-MAY-2013. The calendardatekey column is the one which is used in place of transaction date (each day has a unique calendardatekey). The query runs for couple of hours and breaks. When I contacted DBA they told that its trying to do a full table scan and pulling huge amount of data and doing a kind of sort and then breaks. Could someone please help if there is a better approach? If it works out for one day I will have to perform this for few more days.
Oracle version - 11g
trans table :
columns customerkey, programid and amount - are indexed.
card table :
column customerkey, status are indexed.
select t.customerkey, t.programid, sum(t.amount), max(calendardatekey) from trans t
where
exists (select null from card c where t.customerkey = c.customerkey and c.status = 'A')
--and t.programid = 10
group by t.customerkey, t.instprogkey
having sum(t.amount) < -100
and max(calendardatekey) =4888
Thanks for your help!
"and breaks" is totally devoid of actionable information. Surely you got an actual error message (ora-nnnnn) that you could share with us?
Sounds like your DBA is trying to avoid providing any assistance .....

Server goes into constant Full GC loop after running for a day

Similar Messages

Maybe you are looking for