Server hangs or freezes during heavy load

During peak times of the day, especially during heavy load on the Calendar Server,
the application seems to hang. The client side application will not respond on
the user's desktop, and uni* commands on the server itself respond considerably
slow.
<P>
There are two parameters in the server configuration file that are strongly
believed to be a trigger of server hangs or freezes in large deployments and/or
busy servers. Here is a description of the problem:
<P>
Large deployments tend to be 3000+ users per node. This could be a single or
multi-node environment.
<P>
A lock manager fix was implemented in 4.0 to correct a problem that was
found in 3.51 where the server would hang. At that time, the parameters called
read/writelocktimeouts
were introduced as a failover mechanism in case the
database was not available, which would then trigger the client process to
disconnect rather than hang the whole server.
<P>
These timeouts effectively will terminate a process whose read or write exceeds
the specified periods. The default of 20 seconds is quite a large amount of time;
however, it is not totally unlikely that such a value could be met on a
very busy system. If this is the case, and there is some relation between a
process being terminated by one of these timeouts and subsequent system
instability, then the "solution" would not be to extend the values of the
timeouts but rather to exclude them. This way, it will ensure that no process is
terminated this way and therefore the process would be allowed to continue until
it had completed its job.
<P>
The timeouts were not removed from the product, but under normal circumstances
they probably won't be needed anymore anyhow. It seems that on a busy calendar
server, setting the db timeout alarms may actually trigger the server to freeze.
Below are some examples of errors that appear in the log files which show
that the database is no longer accepting client requests:
<P>
db_VISTA ERROR -920 -> cst_d_open: d_open
db_SchedBaseOpen: unable to open database
probable cause: unilckd is down or "/users/unison/tmp/unisonlckm"
was removed
uniengd: database lock timeout
ITEM: "NA,NA" <0,0>
CLIENT: "unises", "A.02.80"
INET-NAME:
INET-ADDR:
CALL: "SessionsInfoGet"
<P>
To make the fix:
<OL>
<LI>Using your favorite editor, edit the /users/unison/misc/unison.ini file.
In the following section you will see these two parameters:
<P>
[ENG]
writelocktimeout = 20
readlocktimeout = 20
<P>
<LI>Place a "#" sign (or the appropriate comment symbol for your OS) in front of
these two lines and save the file.
<P>
<LI>The server will now have to be restarted in order for the changes to take
effect.
</OL>

This looks similar to what I'm seeing.
DPM 2010, there's one backup set (for me a file server disk) that every time I try to run the initial replica on it the server hangs and needs to be rebooted by iLO. It doesn't just die suddenly, first the data stream on the backup stops then the OS becomes
less responsive but there is no resource issue. trying to open event view will cause a few things to lock up then over a few mins the server is complete froze. like the disk drives have been locked.
Suspecting McAfee, I added in all the exclusions, that didn't help so I added the process exclusions which are done by setting dpmra and csc to low risk and that didn't help either. I could reproduce it just by kicking off a backup for this one file servers
drive so it's easy to test with.
Tonight, I had some permissions in EPO to let me stop the scanning completely and disable the on-access scan and for the first time it worked!
There is definitely an issue between DPM and McAfee beyond what is on MS's web page for AV checks.
I don't have a workaround yet other than stopping the AV completely... Something to follow up on next week. For the moment I made some progress though.

Similar Messages

  • IBM OS ISSUE: SYSTEM HANG WITH JFS2 AND HEAVY LOAD

    제품 : ORACLE SERVER
    작성날짜 : 2004-11-24
    IBM OS ISSUE: SYSTEM HANG WITH JFS2 AND HEAVY LOAD
    ==================================================
    PURPOSE
    이 문서는 IBM AIX5L OS 상에서 Arch Process가 online redo log
    를 release하지 못하고 OS의 kill 명령으로 kill조차 되지 않을 경우
    조치 방법이다.
    Problem Description
    이 문서는 다음 OS 정보와 Oracle version인 경우에 해당한다.
    Oracle Server - Enterprise Edition - Version: 9.2.0.5.0
    Oracle Server - Enterprise Edition - Version: 8.1.7.4.0
    AIX5L Based Systems (64-bit) or (32-bit)
    문제 발생 당시의 증상은 다음과 같다.
    - arch-process doesn't release redolog files
    - arch-process doesn't die after database shutdown
    - arch-process can't be killed
    genclntsh 수행 시 hang 상태이다.
    (sleeping...)28590: kwaitpid(0x2FF22470, -1, 4, 0x00000000, 0x00000000)
    또는 이 문제로 인하여 ORA-600[2103] 에러가 발생하기도 한다.
    이 에러가 발생하는 원인은 IBM AIX5L 64bit, 32bit OS 이슈에 있다.
    IBM OS issue: SYSTEM HANG WITH JFS2 AND HEAVY LOAD.
    Workaround
    none
    Solution Description
    OS AIX5L ML4.0(5.2.0.40)의 bos.up / bos.mp / bos.mp64 fileset level을
    갖는 JFS2 file system을 갖는 경우에 해당한다.
    AIX 5.3 APAR IY59387은 IY58143에 포함되어 있다.
    JFS2 file system에 Heavy I/O activity가 발생하면 database process에
    HANG 문제를 야기시킬 수 있다. 이 때 database process가 kill -9
    명령으로 kill되지 않는다. 데이타베이스를 SHUTDOWN하여도 Process가
    계속 살아 있게 된다.
    Process에게 SIGKILL 함수가 issue되었지만 영향을 주지 않게 된다.
    Process를 제거하는 유일한 방법은 OS를 reboot하는 것이다.
    문제를 예방하는 방법은 IBM APAR-FIX IY59082 를 적용하는 것이다.
    이 APAR는 bos.64bit 5.2.0.42에 포함되어 있다.
    Fix :
    Need to install the IBM APAR-FIX IY59082, this APAR is included on
    bos.64bit 5.2.0.42.
    (주) IBM 엔지니어에게 확인 결과 이 Patch 안에 AIX5L 32bit 용과 64bit
    용이 모두 포함되어 있는 것으로 확인됨.
    Reference Documents
    <Note:285158.1>

    b,
    The install was from my original Dell XP Pro-SP1 'Reinstallation' CD along with the complete MS XP-SP2 download that I burned to another CD. I also had all the latest Dell drivers from their website burned to a CD. Install went as follows:
    - Format HD
    - Install XP Pro SP1
    - Install SP2
    - Install Dell Drivers
    - Obtain any MS patches from their site
    - Install Office
    - Update Office from the MS website
    - Install iTunes
    - Tweak MS settings
    - Install many software programs
    - Make an interim Image of system and store to ExHD
    - Install other software programs (mostly security programs)
    - Make another 'complete' Image to ExHD
    - Restore backup data and music
    After additional testing last night, I determined that it is not iTunes at all (or to blame). I get intermittent hangs/freezes/pauses in other media players (WMP, etc.).
    Seems that something is causing the 'System' process to periodically jump up in CPU %age (though not more than 30-50%). Even this relatively small amount is enough to pause the playback of music or video, hang the mouse movement for a brief moment, or delay a menu opening. All very subtle - you have to be looking for it to notice (except with real-time audio-video, which is quite noticeable whenever this happens).
    I had this problem a year ago, which is why I reimaged then and stayed with SP1 on the advice of Dell and Webroot. I'm stumped as to where to look for the culprit. Every time I go back to the original HD with SP1, everything runs smoothly. Some sort of system call is generating this complete pause....
    I'll be limited in my Internet ability for the next several days, so I may be out of touch until Friday. I'll keep testing and playing with the limited equipment I take along....
    Any thought you or others want to toss out, I'll attempt when possible.
    Thanks!

  • Why my mouse pointer often freeze during heavy hard disk activity?

    Hi!
    My mouse pointer often freeze during heavy disk activity: copying files, extracting archives, downloading and encrypting files from mega.co.nz, etc. The whole system becomes unresponsive. Currently I am using linux-pf kernel with BFS scheduler and I/O scheduler elevator BFQ. What can I do to avoid this annoyance?

    Konstantin_hu wrote:Only 200 MB of swap space of 4 GB is in use.
    Swap space is extremely slow, so you don't want it to be used except as a last resort. Try in /etc/sysctl.conf:
    vm.swappiness=0

  • Server 2008 sp2 freezes during DPM 2010 volume shadow backup

    Hi All,
    We have a hyperv guest server 2008 sp2 that freezes during DPM 2010 volume shadow backup.
    I presume this is when backing up SQL databases. There are no errors in the event logs.
    The sequense of entries in the System event are as follows up until the server freezes.
    1) The DPMRA service entered the running state.
    2) The Volume Shadow Copy service entered the running state. 
    3) DCOM  started the service swprv with arguments "" in order to run the server:
    {65EE1DBA-8FF4-4A58-AC1C-3470EE2F376A}
    4) The Microsoft Software Shadow Copy Provider service entered the running state.
    After this the entries stop and the new entries are from after reboot.                                                               
    You cannot send control-alt-delete or connect to the server in any way.
    Only hard reboot gets it going again.This is the only server this is happening to.
    Please advise if anybody has experienced this and how they resolved.
    Maybe I require a Hotfix.

    This looks similar to what I'm seeing.
    DPM 2010, there's one backup set (for me a file server disk) that every time I try to run the initial replica on it the server hangs and needs to be rebooted by iLO. It doesn't just die suddenly, first the data stream on the backup stops then the OS becomes
    less responsive but there is no resource issue. trying to open event view will cause a few things to lock up then over a few mins the server is complete froze. like the disk drives have been locked.
    Suspecting McAfee, I added in all the exclusions, that didn't help so I added the process exclusions which are done by setting dpmra and csc to low risk and that didn't help either. I could reproduce it just by kicking off a backup for this one file servers
    drive so it's easy to test with.
    Tonight, I had some permissions in EPO to let me stop the scanning completely and disable the on-access scan and for the first time it worked!
    There is definitely an issue between DPM and McAfee beyond what is on MS's web page for AV checks.
    I don't have a workaround yet other than stopping the AV completely... Something to follow up on next week. For the moment I made some progress though.

  • How to redirect to a server busy page when under heavy load

    Hello,
              I have been doing extensive load testing of a web application using Weblogic
              5.10 sp 06. I have found a point where under extremly heavy load the server
              just does not respond anymore. Fair enough.
              What I want to do is at a certain load level (before it stops responding) I
              want to redirect users to a "Server Busy - try again later" page. Is there a
              setting in weblogic that allows me to do this ? Or do I need to have other
              monitoring software to take care of this ?
              Thanks and Regards,
              Nick H
              

    Cameron Purdy <[email protected]> wrote:
              > Hi Nick,
              > Unfortunately, last I check there was no such processing. Weblogic maintains
              > a big (2^16) queue that it puts requests into and (if I remember correctly)
              > it doesn't gracefully handle overflow. I believe the architecture should
              > have been a smaller queue with the overflow condition being
              > protocol-specific (such as HTTP doing a "too busy" error).
              It is possible to create and use your own execute queue in 6.1 and specify it's
              length, so I expected this to happen when queue length reaches this number, but
              it didn't.
              > Peace,
              > --
              > Cameron Purdy
              > Tangosol, Inc.
              > http://www.tangosol.com
              > Tangosol: How Weblogic applications are customized
              > "NH" <[email protected]> wrote in message
              > news:[email protected]...
              >> Hello,
              >>
              >> I have been doing extensive load testing of a web application using
              > Weblogic
              >> 5.10 sp 06. I have found a point where under extremly heavy load the
              > server
              >> just does not respond anymore. Fair enough.
              >>
              >> What I want to do is at a certain load level (before it stops responding)
              > I
              >> want to redirect users to a "Server Busy - try again later" page. Is there
              > a
              >> setting in weblogic that allows me to do this ? Or do I need to have other
              >> monitoring software to take care of this ?
              >>
              >> Thanks and Regards,
              >>
              >> Nick H
              >>
              >>
              Dimitri
              

  • Safari causes system freeze during page load

    Description:
    I've been trying to figure this one out for a while now (about 2 weeks), without any luck. I've tried reading a couple of threads describing similar problems, but it looks like I've got something else going on here seeing as it affect the entire system. I appreciate any suggestion!
    When a window in Safari is open and you're loading a page, it sometimes (almost always) causes a system freeze. Any action performed on any input device within OSX while a page is loading gets delayed by 5-20 seconds. Anything from right-clicking, to opening a new tab in Safari, to navigating to a new folder in Finder.... literally anything. The actual loading of the page is also affected by this. So when I click a link, the page becomes white and the system stalls. After X seconds, the system resumes and the page is displayed fully loaded. Once the page is loaded, everything is back to normal. I can just leave the Safari-window open without anything happening. It's just during the loading of a page.
    Using the mouse to move and switch between open windows are the only actions I seem to be able to perform during the system freeze.
    When it doesn't cause a system-wide freeze, it will still take forever for any page to load. Which isn't a problem in any other browser and I've never had any speed issues in general on any device.
    What I know for sure is that I can solve the problem by closing all open Safari-tabs and windows or closing the app completely, and use Chrome instead. But I do like Safari, so I hope there's another way to fix this.
    I should also note that I've made no changes to the system for many weeks now, other than installing whatever updates might have been available on the App Store.
    What I've tried within Safari:
    - program restart
    - reset settings (every option checked)
    - disable and remove extensions (only had one installed, LiveReload for Adobe Brackets)
    - empty cache
    - delete cookies
    What I've tried within OSX:
    - system restart
    - complete OSX-recovery
    - activity monitor (CPU, HDD and RAM always within normal values during all of this)
    - checked HDD (successfull)
    - checked internet-connection (got 50Mbit/s connection, no speed issues, checked on multiple devices)
    - tried other browsers (Chrome (OSX) and IE (Win 7, Parallels) both run smoothly)
    - turned off IPv6 in terminal (based on a suggestion here at the forums)
    - disabled iCloud-sync
    - disabled Dropbox
    - probably a couple of other things i can't remember right now

    So... I finally fixed it.
    In addition to the things I mentioned I tried, I also checked if it could be malware-related. It wasnt. I then checked if it was just affecting my user-account or OSX globally. I created a new account and Safari ran perfectly. I therefore assumed it had to be something in my library-folder.
    I deleted just about any mention of Safari in the Library-folder (the visible one) and just to be absolutely sure, I deleted the Contents-folder inside Safari.app in the Applications-folder. I then downloaded and installed Safari 7.0.3. Everything runs smooth now.

  • Server freeze during unload - load of a cube(urgetn!!!)

    Hi guys, Occasionally there is a hung up of the server when either the scheduler or an admin tried to unload database and load it back again. Can anyone suggest the possible problems?? Also, wanted to confirm that does the unload of the database , logoff all the users connected to it or in anyways impact users just logged on and not using any db. Please let me know, its urgent.Thanks.

    So... I finally fixed it.
    In addition to the things I mentioned I tried, I also checked if it could be malware-related. It wasnt. I then checked if it was just affecting my user-account or OSX globally. I created a new account and Safari ran perfectly. I therefore assumed it had to be something in my library-folder.
    I deleted just about any mention of Safari in the Library-folder (the visible one) and just to be absolutely sure, I deleted the Contents-folder inside Safari.app in the Applications-folder. I then downloaded and installed Safari 7.0.3. Everything runs smooth now.

  • Daqpad 6508 freezes during heavy usage

    I am having a problem running a daqpad 6508 in a heavy i/o envirnment. This problem seems to be platform dependant, and I'm wondering if there is some way I can slow the Daqpad's usb timing so it doesn't freeze.
    The problem is this: when I run an i/o heavy application (proprietary software using the latest Nidaq drivers) using the 6508 on my 1.3 ghz Dell (win2k) I have no significant problems. However, when I run the same application with the same OS on an embedded platform, the application freezes consistently. The embedded platform is a 500 mhz K6 running on a ALI1541 chipset. Is it possible that the 500 mhz is just too slow to adequately handle USB timings of this nature? or is there some way I can adjust my device so it
    runs slower?
    Thanks,
    --sa

    When you say "handle USB timings of this nature" and "heavy i/o" what are we referring to? How heavy is heavy I/O? Please give me some background on your program so I may determine just how CPU and/or bus intensive it may be.
    The DAQPad-6508 is compatible with the USB 1.1 specification which states that maximum speeds are 1.5 Mbytes/sec. This is slower than PCI and other available busses.
    The issue may also center around the USB driver for your embedded platform. Do you have any other USB devices that you could test out on this OS to verify that only the DAQPad is susceptible to the freezing problem you are seeing.
    Furthermore, does the application generate an error or just lock up? What language are you programming in? Could you troubleshoot wh
    ere the freeze occurs? During which calls to the NI-DAQ driver?
    Regards,
    Justin Britten
    Applications Engineer
    National Instruments

  • Browser freezes during page loading and then either shows the cause as silverlight, or flash to be the cause, on two different computers, since last 3 upgrades

    Since the last three upgrades of Firefox the browser starts to load a web page then freezes. A pop up states that either silverlight or flash have stopped working. This can last for as long as two minutes before it continues to load, or for as little as five seconds.
    I have tried various methods to cure this, by reinstalling silverlight, flash, and Firefox.
    The only permanent cure is to revert back to an older version of Firefox. I would continue with an earlier version but I get sick and tired of popups from the likes of Yahoo, and yourself telling me to upgrade.
    I have also put the the program back to basics, and this is a pain because when I find it's still the same I have to reload all the add ons. Maybe you could devise a method where we only need to click a button to put everything back as it was before resetting it to default.
    This problem is occurring on two totally different computers, both running Intel gear, one i7 at 3.5, with 8 meg of ram, one core 2 duo with 4meg of ram, no drain from graphics cards as they all have their own ram.
    i7 desktop Windows 7 Ultimate 64bit, and Dell XPS Laptop core 2 duo Windows 7 Ultimate 32bit both latest windows 7 versions

    Create a new profile as a test to check if your current profile is causing the problems.
    See "Creating a profile":
    *https://support.mozilla.org/kb/profile-manager-create-and-remove-firefox-profiles
    *http://kb.mozillazine.org/Standard_diagnostic_-_Firefox#Profile_issues
    If the new profile works then you can transfer some files from an existing profile to the new profile, but be cautious not to copy corrupted files to avoid carrying over the problem
    *http://kb.mozillazine.org/Transferring_data_to_a_new_profile_-_Firefox
    Try to disable hardware acceleration in Firefox (you need to close and restart Firefox).
    *Tools > Options > Advanced > General > Browsing: "Use hardware acceleration when available"
    *https://support.mozilla.org/kb/Troubleshooting+extensions+and+themes
    Boot the computer in Windows Safe Mode with network support (press F8 on the boot screen) as a test to see if that helps.

  • Airport (Snow) freezes during heavy traffic

    Anyone else have a problem with their airport base station freezing up during prolonged data transmission? I find that after a couple hours, the base station is totally unresponsive and needs to be reset the hard way (unplugging the power supply). Any suggestions or do i need to buy a new base station?
    MacBook Pro 2GHz   Mac OS X (10.4.6)   2GB Ram

    I too have been having problems lately with the Airport/DSL set up. This roughly corresponded to upgrading our two notebooks to 10.4. For a while, all works well, then the Airport is unresponsive, needs restarting (sometimes that doesn't help either). I can be right next to the sation, and all of a sudden loose one of the signal bars. I belive the problem is with the Airport and not with the ISP becasue when I connect directly to the DSL modem with ethernet (as I am doing now), I have no problem connecting to the internet.
    Occasionally when I turn my airport on and click on the network name I get a message says there was a problem joining the airport network.
    I wonder if any of this has anything to do with 10.4?
    PBG4   Mac OS X (10.4.6)  
    PBG4   Mac OS X (10.4.6)  

  • Trying to down load itunes but it keeps freezing during down load :(

    I cant download itunes for my ipod i was given by a friend ? im new to it all ? also while trying to download itunes it always freezes on me???

    I would suggest giving us more details and clarify both questions.
    From your message, I got these two questions, please confirm that they are the right ones:
    1) You were given an iPod from a friend and you want to load your music onto it?
    2) When you're downloading the iTunes program from the internet, your browser becomes unresponsive and freezes?
    If those aren't the correct questions, please take time to write them in a way that we'll understand exactly what you're trying to do.
    Once we have an idea of what you're asking, we can try to answer properly.
    Thanks!

  • Safari "Freezing" during page loading

    I have never had this problem before I downloaded Safari 4.0 in a stupid software update. While loading a page, it freezes. Safari doesn't lock up, I can still use the menus and can minimize and close the window, but the page does not show on my screen. I can access Google and these Mac forums along with a few other sites. One of which I cannot access is YouTube. It does this every time I try and load a page there. I can't think of what would cause this to happen, but it must be something to do with Safari 4. I have seen other reports of Safari freezing and crashing, but I haven't seen some that seem to be the problem I have. Is there someway to fix it on my own? Because it would be awesome, otherwise I'm just going to have to use Firefox until theres an update for it.
    I even took a video of what happens when I try and access YouTube, since I know sometimes my descriptions of things can be somewhat confusing.
    http://www.youtube.com/watch?v=ozgOjs2NqN0
    Message was edited by: North Kozar

    If you are having the same problem as me, try this:
    http://discussions.apple.com/thread.jspa?threadID=2059943&tstart=0
    It worked perfectly for me and now Safari runs just fine.

  • Inconsistent FileReader behaviour during heavy load

    I am using a StreamTokenizer / FileReader to parse the contents of files and have noticed inconsistent results similar to the ones we get during concurrency problems. Only in my case its a single single Thread calling the parser method.
    I have run the parser 672 times over the same file and I have a counter that counts the parsed words. Though I would expect the counter to always show the same number, suprisingly it sometimes shows a smaller number !
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[190] <====
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[227]
    ---1---[227]
    I have the feeling StreamTokenizer.TT_EOL may return EOF prematurely. Then again I need a deterministic way of processing the contents of as many files as needed. How can I make sure my code will always process the same number of tokens for the same file at any given time?
    Here the method that produces the inconsistent counts:
        public HashMap<String,ArrayList<Long>> parseDocument( Reader       _reader
                                                            , int          _pivDocID
                                                            , yxStopList   _yxStpLst
                                                            , String       _filename )
                throws   FileNotFoundException
                       , IOException
            HashMap<String,ArrayList<Long>> _postingLists = new HashMap<String,ArrayList<Long>>(HASHMAP_INITIAL_SIZE);
            ArrayList<Long>                 _offsets      = null;
            int                             _currOffset   = 0;
            int                             _numericTokens= 0;
            String                          _token        = null;
            BufferedReader                  _buffReader   = new BufferedReader(_reader);
            StreamTokenizer                 _st           = new StreamTokenizer(_buffReader);
            _st.resetSyntax();
            _st.ordinaryChars(0,255);
            _st.eolIsSignificant(true);
            _st.lowerCaseMode(true);
            _st.whitespaceChars(',', ',' ); // COMMA
            _st.whitespaceChars(' ', ' ' ); // SPACE
            _st.whitespaceChars('.', '.' ); // PERIOD
            _st.whitespaceChars('\t','\t'); // TAB
            _st.whitespaceChars('\n','\n'); // EOL
            _st.whitespaceChars('\r','\r'); // EOL
            _st.wordChars('a','z');
            _st.wordChars('A','Z');
            _st.wordChars('0','9');
            _st.wordChars('_','_');
    scan:
            while(true)
                try
                    switch(_st.nextToken())
                            case StreamTokenizer.TT_WORD  :
                                 _token = _st.sval;
                                 if (_token.length() < MINIMUM_ACCEPTABLE_TOKEN_LENGTH) break;
                                 if (_token.length() > MAXIMUM_ACCEPTABLE_TOKEN_LENGTH) break;
                                 if (_token.matches(".*[^a-zA-Z0-9_].*")) break;
                                 if (_token.indexOf("__") > -1) break;
                                 if (_token.matches(".*[0-9].*") && _token.matches(".*[a-zA-Z_].*")) break;
                                 if (!_token.matches("[a-zA-Z_]+"))
                                     if ( !_token.startsWith("0" ) )               break;    //only numbers like: 069 456456 or 004916099113815
                                     if (  _token.startsWith("000") )              break;
                                     if (    _token.length() == 4
                                          && !(    _token.startsWith("19")
                                                || _token.startsWith("20")
                                                || _token.startsWith("21")) )      break;
                                     if (_token.length() > 20)                     break;
                                     _numericTokens++;
                                     if (_numericTokens > MAXIMUM_NUMBERS_PER_DOC) break;    //do not allow too many numeric tokens per document
                                 }//end of [IF]
                                 if ( _yxStpLst.isStopWord(_token) )
                                     yxL.log(6,"[yxParser  --  parseDocument(2)]","","WARNING","REJECTING STOPWORD ["+_token+"] !");
                                     break;
                                 _currOffset++;
                                 if ( !_postingLists.containsKey(_token) )
                                    _offsets = new ArrayList<Long>();
                                    _offsets.add(0,(long)_pivDocID);
                                    _offsets.add(1,(long)_currOffset);
                                    try {
                                       _postingLists.put(_token,_offsets);
                                    }catch(OutOfMemoryError e001){
                                       e001.printStackTrace();
                                       yxL.log(2,"[yxParser  --  parseDocument(2)]","","ERROR","OutOfMemory while parsing ["+_filename+"]");
                                       break scan;
                                 }//end of [IF]
                                 else
                                    _offsets = _postingLists.get(_token);
                                    if (_offsets.size() == MAX_OFFSETS_PER_TOKEN) break;
                                    int _prevSumOfOffsets = 0;
                                    for ( int i1 = 1;                                                       // ignore i1=0 because i1=0 is the DOCID
                                          i1<_offsets.size();                                               // loop until end of encoded Offsets
                                          i1++ ) _prevSumOfOffsets += _offsets.get(i1);                     // sum all existing encoded offsets
                                    _offsets.add(_offsets.size(),(long)(_currOffset - _prevSumOfOffsets));
                                    try {
                                       _postingLists.put(_token,_offsets);              // put : replaces existing Key
                                    }catch(OutOfMemoryError e001){
                                       e001.printStackTrace();
                                       yxL.log(2,"[yxParser  --  parseDocument(2)]","","ERROR","02 [yxParser] OutOfMemory while parsing ["+_filename+"]");
                                       break scan;
                                    yxL.log(6,"[yxParser  --  parseDocument(2)]","","INFO"
                                             ,"02   Updating  ["+_token+"]["+Arrays.toString(_offsets.toArray())+"]");
                                 break;
                            case StreamTokenizer.TT_NUMBER: break;                  // Numbers will be treated as Strings
                            case StreamTokenizer.TT_EOL   : break;                  // EOL
                            case StreamTokenizer.TT_EOF   : break scan;             // EOF
                            default                       :
                                 break;                                             // individual 1-char tokens will be ignored
                    }//end of [SWITCH]
                }catch (Exception e){e.printStackTrace();}
            }//end of [WHILE]
            _buffReader.close();
            _reader.close();
            _token      = null;
            _st         = null;
            _buffReader = null;
            _offsets    = null;
            int _tokensFound = _postingLists.size();
            if (_tokensFound < 1)
                yxL.log(3,"[yxParser  --  parseDocument(2)]","","WARNING","Number of tokens found = ["+_postingLists.size()+"]["+_filename+"]");
                if (sh.isIndexable(_filename))
                    _offsets = new ArrayList<Long>();
                    _offsets.add(0,(long)_pivDocID);
                    _offsets.add(1,1L);
                    try {
                       _postingLists.put(sh.MANUAL_INDEX_FILE_START,_offsets);
                    }catch(OutOfMemoryError e001){
                       e001.printStackTrace();
                       yxL.log(2,"[yxParser  --  parseDocument(2)]","","ERROR","OutOfMemory2 while parsing ["+_filename+"]");
                }//end of [IF]
            else
                yxL.log(4,"[yxParser  --  parseDocument(2)]","","INFO","Number of tokens found = ["+_postingLists.size()+"]");
            COUNT_TOTAL_WORDS += _postingLists.size();
    System.out.println("---1---["+_postingLists.size()+"]");
            yxL.log(6,"[yxParser  --  parseDocument(2)]","END");
            return _postingLists;
        }

    Yes. there is a GUI involved. The Gui participated in this only by triggering the parsing process through the click of a button. The parser receives its input from a Vector which contains fully qualified filenames.
    Each filename is used to instantiate a FileReader Object. This in turn is passed on to the parser method shown above. The inconsistency is located in the method above. I have already searched for days to reach to this conclusion. I was taking for granded that the stream would be read till EOF is reached, but this is not always the case. The method shown above is re-run over 600 times and 2% of the time it return 10-20 words less. !!

  • [(partially?) SOLVED] HP dm1z (4000 model) freezes under heavy load

    Alright, well I've had this laptop for a few weeks now, and it's been pretty good. But I've noticed somewhat random freezes when in Gnome 3 (as I explain below, I've been unable to reproduce this on demand, so I do not know if it would happen were I to use something else or not use X), in which the system is totally unresponsive (pointer doesn't move, unable to switch to a virtual console, no sign of I/O) and has to be turned off manually. The laptop has an AMD E-450 APU, with Radeon HD 6320M graphics built in.
    Magic sysrq stuff sort of works. The last time it happened, I was able to reboot using sysrq, however, I was not able to restore the framebuffer console and thus was unable to get any data from it. And I do know that Alt + printScreen + v works when the system is not in a crashy state.
    The problem has occurred since I got it, and I've kept the system fairly up to date, so the issue has occurred with every kernel version that's been available from three weeks or so ago to now. The problem occurs when using the stock kernel as well as the K8-optimized linux-ck kernel.
    I have read every single log in /var/log, and I have not noticed any messages that seem like they could be associated with the problem. A couple do pop out at me, however:
    Jan 14 10:25:13 localhost logger: ACPI group/action undefined: thermal_zone / LNXTHERM:00
    This has often shown up around the time of the freezes, but also shows up a lot when the system is operating normally. I've got sensord running, however, and the CPU is almost always in the range of 60-70°C, which as far as I can tell is safely below the critical threshold.
    [ 292.154] (II) RADEON(0): radeon_dri2_schedule_flip:670 fevent[0x17eb6d0]
    [ 292.175] (II) RADEON(0): radeon_dri2_flip_event_handler:1067 fevent[0x17eb6d0] width 1366 pitch 5632 (/4 1408)
    Xorg.log gets spammed with this quite a bit, with it showing up anywhere from two or three times a second to once every 30 seconds. Twice so far, I've noticed (after examining Xorg.log.old) post-freeze that only part of the message was written to the log (e.g. just the timestamp, and a handful of characters from the start of the message, or just the first message of the pair).
    Jan 14 10:49:17 localhost kernel: [ 6.170288] [Firmware Bug]: Invalid critical threshold (0)
    This gets written on every boot up.
    I've been unable to find anyone experiencing a similar issue. I am also unable to intentionally reproduce the problem. I've used stress-testing tools to hammer the CPU with at least 3-4 threads, as well as hammer the memory and I/O, but running it for around an hour has not worked.
    Also, it seems to be much more likely to freeze after having just rebooted from a freeze. This suggests to me that it is a thermal issue, but nothing I've seen corroborates that.
    I've really got no idea how to troubleshoot this, so any help at all would be appreciated. And if you need any additional information, just ask. Thanks in advance.
    In case it helps:
    rc.conf
    lsmod output
    lspci output
    Last edited by Guff (2012-07-17 16:36:01)

    @dimath: thanks for the link. There was a link there to another bug report, which mentioned some kernel parameters to try out. Didn't help.
    Since I was having issues with getting my wireless to work properly (which I ultimately was not able to do, so just got a tiny USB dongle for the time being), I did a lot of mucking around in the system. Mucking around with many things that I did not understand. So, I figured well maybe I did something stupid that somehow brought this curse down upon me, so last week I did a total reinstall.
    It didn't work.
    I've tried KDE, and the problem persists, so it's not gnome-specific for me. I'm trying the catalyst driver now, but of course it's fairly unusable with gnome3 at the moment, so I'm only testing it in KDE. It seems like it might be okay, because I've been doing my best to try and trigger a crash/freeze.
    Then again, it took a long while for the issue to pop up once I started using KDE. I still haven't quite worked out how to trigger the damn thing.
    It often shows up when building some larger packages, installing packages, and when launching Firefox (I often have an ungodly amount of tabs open, so it is fairly resource-intensive). As mentioned, however, it's not at all consistent.
    dimath, have you tried another DE yet? And if the problem persisted, have you tried catalyst? Obviously, I don't expect you to use it with gnome given that it's buggy as hell, but I'd be interested to see if it helps for you at all.

  • Broken TCP stack in latest kernel when under heavy load

    I'm running an Arch box with a decent amount of HTTP traffic. After upgrading to the latest kernel I've seen that packets are send from the wrong source and destination address. This only applies during heavy load (100+ requests per second). tcpdump shows the following:
    18:52:58.512573 IP 0.0.0.0.80 > 0.0.0.0.4316: Flags [FP.], seq 0, ack 1, win 14400, length 0
    18:52:58.512600 IP 0.0.0.0.80 > 0.0.0.0.56546: Flags [FP.], seq 0, ack 1, win 14400, length 0
    18:52:58.512621 IP 0.0.0.0.80 > 0.0.0.0.4535: Flags [FP.], seq 0, ack 1, win 14600, length 0
    18:52:58.512641 IP 0.0.0.0.80 > 0.0.0.0.3528: Flags [FP.], seq 0, ack 1, win 14600, length 0
    18:52:58.512662 IP 0.0.0.0.80 > 0.0.0.0.4509: Flags [FP.], seq 0, ack 1, win 14400, length 0
    18:52:58.512682 IP 0.0.0.0.80 > 0.0.0.0.65040: Flags [FP.], seq 0, ack 1, win 14600, length 0
    18:52:58.512702 IP 0.0.0.0.80 > 0.0.0.0.2455: Flags [FP.], seq 0, ack 1, win 10240, length 0
    18:52:58.512722 IP 0.0.0.0.80 > 0.0.0.0.16545: Flags [FP.], seq 0:268, ack 1, win 15008, length 268
    18:52:58.519258 IP 0.0.0.0.80 > 0.0.0.0.29802: Flags [FP.], seq 0:268, ack 1, win 980, options [nop,nop,TS val 745514 ecr 1317559555], length 268
    18:52:58.565907 IP 0.0.0.0.80 > 0.0.0.0.32376: Flags [FP.], seq 0, ack 1, win 14400, length 0
    18:52:58.619241 IP 0.0.0.0.80 > 0.0.0.0.50493: Flags [FP.], seq 0:268, ack 1, win 11256, options [nop,nop,TS val 745544 ecr 9539361], length 268
    18:52:58.805927 IP 0.0.0.0.80 > 0.0.0.0.20852: Flags [FP.], seq 3025419976:3025420244, ack 3037671074, win 967, options [nop,nop,TS val 745600 ecr 6445640], length 268
    18:52:58.805953 IP 0.0.0.0.80 > 0.0.0.0.65025: Flags [FP.], seq 1663827778:1663828046, ack 2127675352, win 707, options [nop,nop,TS val 745600 ecr 457812708], length 268
    18:52:58.845918 IP 0.0.0.0.80 > 0.0.0.0.2217: Flags [FP.], seq 0:268, ack 1, win 707, options [nop,nop,TS val 745612 ecr 546643], length 268
    18:52:59.099245 IP 0.0.0.0.80 > 0.0.0.0.5112: Flags [FP.], seq 0:268, ack 1, win 15008, length 268
    18:52:59.152582 IP 0.0.0.0.80 > 0.0.0.0.1175: Flags [FP.], seq 0:268, ack 1, win 15008, length 268
    18:52:59.232612 IP 0.0.0.0.80 > 0.0.0.0.47217: Flags [FP.], seq 684621876:684622144, ack 3544859356, win 11256, length 268
    18:52:59.659258 IP 0.0.0.0.80 > 0.0.0.0.3098: Flags [FP.], seq 2105858244:2105858512, ack 3896053916, win 980, options [nop,nop,TS val 745856 ecr 52041], length 268
    18:52:59.659290 IP 0.0.0.0.80 > 0.0.0.0.3099: Flags [FP.], seq 18772067:18772335, ack 2568646283, win 980, options [nop,nop,TS val 745856 ecr 52041], length 268
    18:52:59.759244 IP 0.0.0.0.80 > 0.0.0.0.18780: Flags [FP.], seq 0:268, ack 1, win 707, options [nop,nop,TS val 745886 ecr 168876], length 268
    18:52:59.845907 IP 0.0.0.0.80 > 0.0.0.0.58449: Flags [FP.], seq 0, ack 1, win 980, options [nop,nop,TS val 745912 ecr 528058426], length 0
    18:52:59.925936 IP 0.0.0.0.80 > 0.0.0.0.65137: Flags [FP.], seq 0:268, ack 1, win 15008, length 268
    18:52:59.979497 IP 0.0.0.0.80 > 0.0.0.0.2920: Flags [FP.], seq 0:268, ack 1, win 980, options [nop,nop,TS val 745952 ecr 18879], length 268
    18:52:59.979527 IP 0.0.0.0.80 > 0.0.0.0.2922: Flags [FP.], seq 0:268, ack 1, win 980, options [nop,nop,TS val 745952 ecr 18879], length 268
    18:52:59.979553 IP 0.0.0.0.80 > 0.0.0.0.2940: Flags [FP.], seq 0:268, ack 1, win 980, options [nop,nop,TS val 745952 ecr 18879], length 268
    Source and destination ports are correctly set. Wireshark shows the correct HTML inside the packets that are returned to 0.0.0.0. The web server log also looks normal; the correct IP address is displayed and logged as a successful request.
    When dropping incomming traffic on port 80 on eth0 everything works as expected (when requesting the server on eth1, which otherwise fails).
    I'm running on "Linux srv 3.0-ARCH #1 SMP PREEMPT Wed Oct 19 12:14:48 UTC 2011 i686" which is the latest kernel in the repos. When booting the fallback image this problem does not exist, all packets are correctly addressed no matter how much load I put on the server.
    Does anyone else have this problem?
    Edit:
    Running lighttpd 1.4.29. No tweaked kernel/TCP parameters whatsoever.
    Last edited by nullvoid (2011-10-29 17:19:57)

    Did a full reinstall of Arch on another machine and the problem still persist. Tried with Apache and Nginx, same behaviour as with Lighttpd. Could anyone else using an arch box under heavy load see if there's activity from 0.0.0.0?
    Hint:
    # tcpdump -n host 0.0.0.0
    I'll do a bug report upstream later today.

Maybe you are looking for