From amolk112k at gmail.com Mon Jul 2 13:45:18 2012 From: amolk112k at gmail.com (Amol Kulkarni) Date: Mon, 2 Jul 2012 19:15:18 +0530 Subject: 100% cpu usage on CentOS/RHEL 6 Message-ID: Hi, I have a java app running on CentOS 6.0. It always runs in background via cron. Sometimes this app goes into wait state while using 100% cpu. My java version is : java version "1.6.0_17" OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) Other symptoms are : a. One thread of the process seems to be in a loop waiting for something. When traced using strace, it shows following o/p continuously : > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 1, {1340347489,> 822867000}, ffffffff) = -1 ETIMEDOUT (Connection timed out) b. It seems like the process has finished working, looking at the files it is using. Only few files are remaining. The output of 'ls /proc/pid/fd/ shows : > lr-x------ 1 root root 64 Jun 22 13:13 0 -> pipe:[77107601] > l-wx------ 1 root root 64 Jun 22 13:13 1 -> pipe:[77120162] > l-wx------ 1 root root 64 Jun 22 13:13 2 -> /var/log/mithi/mcs/agent_account_mailstore_exceed_limit.sh.log > lr-x------ 1 root root 64 Jun 22 13:13 3 -> /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/jre/lib/rt.jar More specifically, are there any known problems in running openjdk based Java processes in background on CentOS/RHEL 6? I'm able to simulate the problem with a very simple script : #!/bin/bash while [ 1 ] do /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/bin/java -version & sleep 1s done When the above script is run for about 3 - 4 hours, one or two java processes go in hang state with cpu usage touching 100%. Has anybody faced similar situation? Any clues or references would be very helpful. From mthornton at optrak.com Mon Jul 2 14:02:14 2012 From: mthornton at optrak.com (Mark Thornton) Date: Mon, 02 Jul 2012 15:02:14 +0100 Subject: 100% cpu usage on CentOS/RHEL 6 In-Reply-To: References: Message-ID: <4FF1A9E6.4040503@optrak.com> On 02/07/12 14:45, Amol Kulkarni wrote: > Hi, > > I have a java app running on CentOS 6.0. It always runs in background via > cron. Sometimes this app goes into wait state while using 100% cpu. > > My java version is : > > java version "1.6.0_17" > OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > Other symptoms are : > > a. One thread of the process seems to be in a loop waiting for something. > When traced using strace, it shows following o/p continuously : > >> futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 >> futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, Could it be related to this: https://bugzilla.kernel.org/show_bug.cgi?id=32922 Mark Thornton From fcassia at gmail.com Mon Jul 2 18:38:44 2012 From: fcassia at gmail.com (Fernando Cassia) Date: Mon, 2 Jul 2012 15:38:44 -0300 Subject: OpenJDK level on CentOS? Message-ID: What level of OpenJDK can one find in the latest version of CentOS?. I've looked at the CentOS wiki pages but couldn't find a precise answer, other than references on how to manually install Sun and IBM JREs, and a mention about OpenJDK 6 being part of the repos. http://wiki.centos.org/HowTos/JavaOnCentOS TIA, FC From aph at redhat.com Mon Jul 2 19:08:33 2012 From: aph at redhat.com (Andrew Haley) Date: Mon, 02 Jul 2012 20:08:33 +0100 Subject: OpenJDK level on CentOS? In-Reply-To: References: Message-ID: <4FF1F1B1.80507@redhat.com> On 07/02/2012 07:38 PM, Fernando Cassia wrote: > What level of OpenJDK can one find in the latest version of CentOS?. > I've looked at the CentOS wiki pages but couldn't find a precise > answer, other than references on how to manually install Sun and IBM > JREs, and a mention about OpenJDK 6 being part of the repos. > > http://wiki.centos.org/HowTos/JavaOnCentOS That sounds like good information. What more do you want to know? Andrew. From fcassia at gmail.com Mon Jul 2 19:22:27 2012 From: fcassia at gmail.com (Fernando Cassia) Date: Mon, 2 Jul 2012 16:22:27 -0300 Subject: OpenJDK level on CentOS? In-Reply-To: <4FF1F1B1.80507@redhat.com> References: <4FF1F1B1.80507@redhat.com> Message-ID: On Mon, Jul 2, 2012 at 4:08 PM, Andrew Haley wrote: > That sounds like good information. What more do you want to > know? The last CentOS version is: x.y and it includes in its repos OpenJDK ver x.y.z. FC From david.holmes at oracle.com Tue Jul 3 02:24:48 2012 From: david.holmes at oracle.com (David Holmes) Date: Tue, 03 Jul 2012 12:24:48 +1000 Subject: 100% cpu usage on CentOS/RHEL 6 In-Reply-To: <4FF1A9E6.4040503@optrak.com> References: <4FF1A9E6.4040503@optrak.com> Message-ID: <4FF257F0.9060403@oracle.com> Yes this sounds like the linux leap-second bug. Despite a lot of ignorant comments on the web this is not a bug in Java. But I would really appreciate it if someone experience this can send me a stack trace from the problematic thread(s). I think I know what is happening but need to confirm. Thanks, David Holmes On 3/07/2012 12:02 AM, Mark Thornton wrote: > On 02/07/12 14:45, Amol Kulkarni wrote: >> Hi, >> >> I have a java app running on CentOS 6.0. It always runs in background via >> cron. Sometimes this app goes into wait state while using 100% cpu. >> >> My java version is : >> >> java version "1.6.0_17" >> OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) >> OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) >> >> Other symptoms are : >> >> a. One thread of the process seems to be in a loop waiting for something. >> When traced using strace, it shows following o/p continuously : >> >>> futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 >>> futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, > > Could it be related to this: > https://bugzilla.kernel.org/show_bug.cgi?id=32922 > > Mark Thornton > From aph at redhat.com Tue Jul 3 07:35:42 2012 From: aph at redhat.com (Andrew Haley) Date: Tue, 03 Jul 2012 08:35:42 +0100 Subject: OpenJDK level on CentOS? In-Reply-To: References: <4FF1F1B1.80507@redhat.com> Message-ID: <4FF2A0CE.7050200@redhat.com> On 07/02/2012 08:22 PM, Fernando Cassia wrote: > On Mon, Jul 2, 2012 at 4:08 PM, Andrew Haley wrote: >> That sounds like good information. What more do you want to >> know? > > The last CentOS version is: x.y > and it includes in its repos OpenJDK ver x.y.z. So why don't you just use yum update? Then you'll know. It's far more useful to ask CentOS about their plans. Andrew. From henri.gomez at gmail.com Tue Jul 3 08:21:33 2012 From: henri.gomez at gmail.com (Henri Gomez) Date: Tue, 3 Jul 2012 10:21:33 +0200 Subject: OpenJDK level on CentOS? In-Reply-To: <4FF2A0CE.7050200@redhat.com> References: <4FF1F1B1.80507@redhat.com> <4FF2A0CE.7050200@redhat.com> Message-ID: yum search openjdk java-1.6.0-openjdk.x86_64 : OpenJDK Runtime Environment java-1.6.0-openjdk-demo.x86_64 : OpenJDK Demos java-1.6.0-openjdk-devel.x86_64 : OpenJDK Development Environment java-1.6.0-openjdk-javadoc.x86_64 : OpenJDK API Documentation java-1.6.0-openjdk-src.x86_64 : OpenJDK Source Bundle icedtea-web.x86_64 : Additional Java components for OpenJDK --- java version "1.6.0_24" OpenJDK Runtime Environment (IcedTea6 1.11.3) (rhel-1.48.1.11.3.el6_2-x86_64) OpenJDK 64-Bit Server VM (build 20.0-b12, mixed mode) 2012/7/3 Andrew Haley : > On 07/02/2012 08:22 PM, Fernando Cassia wrote: >> On Mon, Jul 2, 2012 at 4:08 PM, Andrew Haley wrote: >>> That sounds like good information. What more do you want to >>> know? >> >> The last CentOS version is: x.y >> and it includes in its repos OpenJDK ver x.y.z. > > So why don't you just use yum update? Then you'll know. It's > far more useful to ask CentOS about their plans. > > Andrew. > From amolk112k at gmail.com Tue Jul 3 11:02:52 2012 From: amolk112k at gmail.com (Amol Kulkarni) Date: Tue, 3 Jul 2012 16:32:52 +0530 Subject: 100% cpu usage on CentOS/RHEL 6 In-Reply-To: <4FF1A9E6.4040503@optrak.com> References: <4FF1A9E6.4040503@optrak.com> Message-ID: Hi, >From the symptoms, it looks like the same bug to me too. I'll check whether it is solved in CentOS/RHEL 6 kernels. On Mon, Jul 2, 2012 at 7:32 PM, Mark Thornton wrote: > On 02/07/12 14:45, Amol Kulkarni wrote: > > Hi, > > I have a java app running on CentOS 6.0. It always runs in background via > cron. Sometimes this app goes into wait state while using 100% cpu. > > My java version is : > > java version "1.6.0_17" > OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > Other symptoms are : > > a. One thread of the process seems to be in a loop waiting for something. > When traced using strace, it shows following o/p continuously : > > > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, > > > Could it be related to this: > https://bugzilla.kernel.org/show_bug.cgi?id=32922 > > Mark Thornton > > From amolk112k at gmail.com Tue Jul 3 11:07:04 2012 From: amolk112k at gmail.com (Amol Kulkarni) Date: Tue, 3 Jul 2012 16:37:04 +0530 Subject: 100% cpu usage on CentOS/RHEL 6 In-Reply-To: <4FF257F0.9060403@oracle.com> References: <4FF1A9E6.4040503@optrak.com> <4FF257F0.9060403@oracle.com> Message-ID: I dont think this is the leap-second bug bcos I can simulate it at any time using foll script. #!/bin/bash while [ 1 ] do /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/bin/java -version & sleep 1s done The leap second bug seems to be occuring only near 30 Jun or 1 Jul. I did generate a backtrace with this prb but I dont have openjdk with debug symbols - can u tell me where I can get it? On Tue, Jul 3, 2012 at 7:54 AM, David Holmes wrote: > Yes this sounds like the linux leap-second bug. > > Despite a lot of ignorant comments on the web this is not a bug in Java. > > But I would really appreciate it if someone experience this can send me a > stack trace from the problematic thread(s). I think I know what is > happening but need to confirm. > > Thanks, > David Holmes > > > On 3/07/2012 12:02 AM, Mark Thornton wrote: > >> On 02/07/12 14:45, Amol Kulkarni wrote: >> >>> Hi, >>> >>> I have a java app running on CentOS 6.0. It always runs in background via >>> cron. Sometimes this app goes into wait state while using 100% cpu. >>> >>> My java version is : >>> >>> java version "1.6.0_17" >>> OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) >>> OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) >>> >>> Other symptoms are : >>> >>> a. One thread of the process seems to be in a loop waiting for something. >>> When traced using strace, it shows following o/p continuously : >>> >>> futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 >>>> futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|**FUTEX_CLOCK_REALTIME, >>>> >>> >> Could it be related to this: >> https://bugzilla.kernel.org/**show_bug.cgi?id=32922 >> >> Mark Thornton >> >> From david.holmes at oracle.com Tue Jul 3 11:22:53 2012 From: david.holmes at oracle.com (David Holmes) Date: Tue, 03 Jul 2012 21:22:53 +1000 Subject: 100% cpu usage on CentOS/RHEL 6 In-Reply-To: References: <4FF1A9E6.4040503@optrak.com> <4FF257F0.9060403@oracle.com> Message-ID: <4FF2D60D.1080604@oracle.com> On 3/07/2012 9:07 PM, Amol Kulkarni wrote: > I dont think this is the leap-second bug bcos I can simulate it at any > time using foll script. > #!/bin/bash > > while [ 1 ] > do > /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/bin/java -version & > sleep 1s > done > > > The leap second bug seems to be occuring only near 30 Jun or 1 Jul. That's when it activates but unless you take steps to fix it it remains in your system. The time offset used by the hrtimers remains 1 second out of sync with actual time. > I did generate a backtrace with this prb but I dont have openjdk with > debug symbols - can u tell me where I can get it? No sorry, I don't know about the OpenJDK distros available for CentOS. Further discussion should move to hotspot-dev at openjdk.java.net cc'd. David ----- > > On Tue, Jul 3, 2012 at 7:54 AM, David Holmes > wrote: > > Yes this sounds like the linux leap-second bug. > > Despite a lot of ignorant comments on the web this is not a bug in Java. > > But I would really appreciate it if someone experience this can send > me a stack trace from the problematic thread(s). I think I know what > is happening but need to confirm. > > Thanks, > David Holmes > > > On 3/07/2012 12:02 AM, Mark Thornton wrote: > > On 02/07/12 14:45, Amol Kulkarni wrote: > > Hi, > > I have a java app running on CentOS 6.0. It always runs in > background via > cron. Sometimes this app goes into wait state while using > 100% cpu. > > My java version is : > > java version "1.6.0_17" > OpenJDK Runtime Environment (IcedTea6 1.7.4) > (rhel-1.21.b17.el6-x86_64) > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > Other symptoms are : > > a. One thread of the process seems to be in a loop waiting > for something. > When traced using strace, it shows following o/p continuously : > > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > futex(0x7fb8000ac754, > FUTEX_WAIT_BITSET_PRIVATE|__FUTEX_CLOCK_REALTIME, > > > Could it be related to this: > https://bugzilla.kernel.org/__show_bug.cgi?id=32922 > > > Mark Thornton > > From xieliang at xiaomi.com Tue Jul 3 11:24:13 2012 From: xieliang at xiaomi.com (=?gb2312?B?0LvBvA==?=) Date: Tue, 3 Jul 2012 11:24:13 +0000 Subject: =?gb2312?B?tPC4tDogMTAwJSBjcHUgdXNhZ2Ugb24gQ2VudE9TL1JIRUwgNg==?= In-Reply-To: References: <4FF1A9E6.4040503@optrak.com>, Message-ID: <6DECD2A517016948873E76418A593E7009A026C9@EXCHANGE-DAG2.xiaomi.net> May be you can have an eye on here : https://lkml.org/lkml/2012/7/1/27 :) ________________________________________ ???: discuss-bounces at openjdk.java.net [discuss-bounces at openjdk.java.net] ?? Amol Kulkarni [amolk112k at gmail.com] ????: 2012?7?3? 19:02 ?: Mark Thornton Cc: discuss at openjdk.java.net ??: Re: 100% cpu usage on CentOS/RHEL 6 Hi, >From the symptoms, it looks like the same bug to me too. I'll check whether it is solved in CentOS/RHEL 6 kernels. On Mon, Jul 2, 2012 at 7:32 PM, Mark Thornton wrote: > On 02/07/12 14:45, Amol Kulkarni wrote: > > Hi, > > I have a java app running on CentOS 6.0. It always runs in background via > cron. Sometimes this app goes into wait state while using 100% cpu. > > My java version is : > > java version "1.6.0_17" > OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > Other symptoms are : > > a. One thread of the process seems to be in a loop waiting for something. > When traced using strace, it shows following o/p continuously : > > > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, > > > Could it be related to this: > https://bugzilla.kernel.org/show_bug.cgi?id=32922 > > Mark Thornton > > From fcassia at gmail.com Tue Jul 3 11:46:48 2012 From: fcassia at gmail.com (Fernando Cassia) Date: Tue, 3 Jul 2012 08:46:48 -0300 Subject: OpenJDK level on CentOS? In-Reply-To: <4FF2A0CE.7050200@redhat.com> References: <4FF1F1B1.80507@redhat.com> <4FF2A0CE.7050200@redhat.com> Message-ID: On Tue, Jul 3, 2012 at 4:35 AM, Andrew Haley wrote: > So why don't you just use yum update? Then you'll know. It's > far more useful to ask CentOS about their plans. Gee, because I'm NOT running CentOS, just PLAN TO run it, but first I want to know the level of OpenJDK available on it. FC -- During times of Universal Deceit, telling the truth becomes a revolutionary act - George Orwell From roman at kennke.org Tue Jul 3 11:56:52 2012 From: roman at kennke.org (Roman Kennke) Date: Tue, 03 Jul 2012 13:56:52 +0200 Subject: OpenJDK level on CentOS? In-Reply-To: References: <4FF1F1B1.80507@redhat.com> <4FF2A0CE.7050200@redhat.com> Message-ID: <4FF2DE04.4020107@kennke.org> Am 03.07.2012 13:46, schrieb Fernando Cassia: > On Tue, Jul 3, 2012 at 4:35 AM, Andrew Haley wrote: >> So why don't you just use yum update? Then you'll know. It's >> far more useful to ask CentOS about their plans. > > Gee, because I'm NOT running CentOS, just PLAN TO run it, but first I > want to know the level of OpenJDK available on it. From what I can see on: http://mirror.centos.org/centos/6.2/os/x86_64/Packages/ The latest is: java-1.6.0-openjdk-1.6.0.0-1.41.1.10.4.el6.x86_64.rpm Regards, Roman From fcassia at gmail.com Tue Jul 3 12:18:45 2012 From: fcassia at gmail.com (Fernando Cassia) Date: Tue, 3 Jul 2012 09:18:45 -0300 Subject: OpenJDK level on CentOS? In-Reply-To: <4FF2DE04.4020107@kennke.org> References: <4FF1F1B1.80507@redhat.com> <4FF2A0CE.7050200@redhat.com> <4FF2DE04.4020107@kennke.org> Message-ID: On Tue, Jul 3, 2012 at 8:56 AM, Roman Kennke wrote: > From what I can see on: > > http://mirror.centos.org/centos/6.2/os/x86_64/Packages/ > > The latest is: java-1.6.0-openjdk-1.6.0.0-1.41.1.10.4.el6.x86_64.rpm > > Regards, > Roman Thanks Roman. And to Andrew, you can see below that I'm not alone asking about the supported OpenJDK Java release in CentOS... https://www.centos.org/modules/newbb/viewtopic.php?topic_id=37906&forum=56 "After installing CentOS 6.2, I searched the default repositories and found that OpenJDK 6 was there, but OpenJDK 7 was not. Is there another repository that has the RPM's for OpenJDK 7? Are there plans to eventually add it to the default repositories? In the meantime, is it possible to use the RPMs from another distro such as Fedora? I would like to avoid having to compile from source." and the reply: "RHEL 6.3 was released about 2 days ago and one of the features listed is that it includes OpenJDK 7 as well as the Oracle and IBM JVMs. It'll take a while to build CentOS 6.3 from the sources but it should be available soonish." I will now ask on the CentOS mailing list and see if I can edit the wiki to reflect the above info. Thanks to everyone who replied FC From amolk112k at gmail.com Tue Jul 3 13:54:33 2012 From: amolk112k at gmail.com (Amol Kulkarni) Date: Tue, 3 Jul 2012 19:24:33 +0530 Subject: =?GB2312?B?UmU6ILTwuLQ6IDEwMCUgY3B1IHVzYWdlIG9uIENlbnRPUy9SSEVMIDY=?= In-Reply-To: <6DECD2A517016948873E76418A593E7009A026C9@EXCHANGE-DAG2.xiaomi.net> References: <4FF1A9E6.4040503@optrak.com> <6DECD2A517016948873E76418A593E7009A026C9@EXCHANGE-DAG2.xiaomi.net> Message-ID: Seems like I've hit the leap second bug itself. But since I'm getting that error for atleast 15 days now i.e from 15 June or so. Is it likely that the leap second bug occurs before the leap second is introduced by the kernel? On Tue, Jul 3, 2012 at 4:54 PM, ?? wrote: > May be you can have an eye on here : https://lkml.org/lkml/2012/7/1/27 :) > ________________________________________ > ???: discuss-bounces at openjdk.java.net [discuss-bounces at openjdk.java.net] > ?? Amol Kulkarni [amolk112k at gmail.com] > ????: 2012?7?3? 19:02 > ?: Mark Thornton > Cc: discuss at openjdk.java.net > ??: Re: 100% cpu usage on CentOS/RHEL 6 > > Hi, > > >From the symptoms, it looks like the same bug to me too. > I'll check whether it is solved in CentOS/RHEL 6 kernels. > > > > > On Mon, Jul 2, 2012 at 7:32 PM, Mark Thornton > wrote: > > > On 02/07/12 14:45, Amol Kulkarni wrote: > > > > Hi, > > > > I have a java app running on CentOS 6.0. It always runs in background via > > cron. Sometimes this app goes into wait state while using 100% cpu. > > > > My java version is : > > > > java version "1.6.0_17" > > OpenJDK Runtime Environment (IcedTea6 1.7.4) > (rhel-1.21.b17.el6-x86_64) > > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > > > Other symptoms are : > > > > a. One thread of the process seems to be in a loop waiting for something. > > When traced using strace, it shows following o/p continuously : > > > > > > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > > futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, > > > > > > Could it be related to this: > > https://bugzilla.kernel.org/show_bug.cgi?id=32922 > > > > Mark Thornton > > > > > From xieliang at xiaomi.com Tue Jul 3 14:05:14 2012 From: xieliang at xiaomi.com (=?gb2312?B?0LvBvA==?=) Date: Tue, 3 Jul 2012 14:05:14 +0000 Subject: =?gb2312?B?tPC4tDogtPC4tDogMTAwJSBjcHUgdXNhZ2Ugb24gQ2VudE9TL1JIRUwgNg==?= In-Reply-To: References: <4FF1A9E6.4040503@optrak.com> <6DECD2A517016948873E76418A593E7009A026C9@EXCHANGE-DAG2.xiaomi.net>, Message-ID: <6DECD2A517016948873E76418A593E7009A02A0B@EXCHANGE-DAG2.xiaomi.net> If so, it's unlikely a leap-second issue. You can figure it out by `dmesg`, the leap-second worked while sth like "Clock: inserting leap second 23:59:60 UTC" be found and ntpd service running:) ________________________________ ???: Amol Kulkarni [amolk112k at gmail.com] ????: 2012?7?3? 21:54 ?: ?? Cc: Mark Thornton; discuss at openjdk.java.net ??: Re: ??: 100% cpu usage on CentOS/RHEL 6 Seems like I've hit the leap second bug itself. But since I'm getting that error for atleast 15 days now i.e from 15 June or so. Is it likely that the leap second bug occurs before the leap second is introduced by the kernel? On Tue, Jul 3, 2012 at 4:54 PM, ?? > wrote: May be you can have an eye on here : https://lkml.org/lkml/2012/7/1/27 :) ________________________________________ ???: discuss-bounces at openjdk.java.net [discuss-bounces at openjdk.java.net] ?? Amol Kulkarni [amolk112k at gmail.com] ????: 2012?7?3? 19:02 ?: Mark Thornton Cc: discuss at openjdk.java.net ??: Re: 100% cpu usage on CentOS/RHEL 6 Hi, >From the symptoms, it looks like the same bug to me too. I'll check whether it is solved in CentOS/RHEL 6 kernels. On Mon, Jul 2, 2012 at 7:32 PM, Mark Thornton > wrote: > On 02/07/12 14:45, Amol Kulkarni wrote: > > Hi, > > I have a java app running on CentOS 6.0. It always runs in background via > cron. Sometimes this app goes into wait state while using 100% cpu. > > My java version is : > > java version "1.6.0_17" > OpenJDK Runtime Environment (IcedTea6 1.7.4) (rhel-1.21.b17.el6-x86_64) > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) > > Other symptoms are : > > a. One thread of the process seems to be in a loop waiting for something. > When traced using strace, it shows following o/p continuously : > > > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 > futex(0x7fb8000ac754, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, > > > Could it be related to this: > https://bugzilla.kernel.org/show_bug.cgi?id=32922 > > Mark Thornton > > From mark at klomp.org Thu Jul 5 21:31:44 2012 From: mark at klomp.org (Mark Wielaard) Date: Thu, 5 Jul 2012 23:31:44 +0200 Subject: Proposed update to the OpenJDK Web Site Terms of Use In-Reply-To: <20120620191048.9B6579AC@eggemoggin.niobe.net> References: <20120620191048.9B6579AC@eggemoggin.niobe.net> Message-ID: <20120705213144.GH3373@toonder.wildebeest.org> Hi Mark, On Wed, Jun 20, 2012 at 12:10:48PM -0700, mark.reinhold at oracle.com wrote: > The primary goal of this update is to allow work on specifications for > Java SE JSRs to take place in the OpenJDK Community, right alongside the > work on their reference implementations. So the goal is to make the export group discussions, specification proposals and comments public alongside the other OpenJDK comminuty interactions so that it is easy to evaluate JSR to be by making GPL implementations to see how they work out in practice. The FAQ implies you should not mix and match comments, evaluation and implementation code on the same list. But I assume it is fine to post something referencing one or the other (e.g. posting something like "I tried implementing the FooBar property for all Baz classes as described in paragraph 42, I won't post the code on this list, but it can be seen over there in my public repo and is available under the GPL, evaluating the included benchmark shows that... etc.") > The Expert Groups for such JSRs > will hold their technical discussions in the open, for all to see. The > specifications they create will continue to be licensed under terms > similar to those used in the past [2], but no click-through agreement > will be required in order to access specification materials. I take this to mean that you don't need to agree to the terms that are incompatible with creating a GPL implementation anymore just to read the now public specifications. That is really, really useful. Thanks. Cheers, Mark From mark.reinhold at oracle.com Thu Jul 5 22:53:55 2012 From: mark.reinhold at oracle.com (mark.reinhold at oracle.com) Date: Thu, 05 Jul 2012 15:53:55 -0700 Subject: Proposed update to the OpenJDK Web Site Terms of Use In-Reply-To: mark@klomp.org; Thu, 05 Jul 2012 23:31:44 +0200; <20120705213144.GH3373@toonder.wildebeest.org> Message-ID: <20120705225355.1BC45888@eggemoggin.niobe.net> 2012/7/5 14:31 -0700, mark at klomp.org: > ... The FAQ implies > you should not mix and match comments, evaluation and implementation > code on the same list. But I assume it is fine to post something > referencing one or the other (e.g. posting something like "I tried > implementing the FooBar property for all Baz classes as described in > paragraph 42, I won't post the code on this list, but it can be seen > over there in my public repo and is available under the GPL, evaluating > the included benchmark shows that... etc.") Yes. IANAL, but I would think that's perfectly fine. >> The Expert Groups for such JSRs >> will hold their technical discussions in the open, for all to see. The >> specifications they create will continue to be licensed under terms >> similar to those used in the past [2], but no click-through agreement >> will be required in order to access specification materials. > > I take this to mean that you don't need to agree to the terms that are > incompatible with creating a GPL implementation anymore just to read > the now public specifications. That is really, really useful. Thanks. You don't need to agree explicitly to the comment-and-evaluation terms by clicking on an "I agree" button. They're still there, though, whether you agree with them or not. - Mark From mark at klomp.org Fri Jul 6 21:26:36 2012 From: mark at klomp.org (Mark Wielaard) Date: Fri, 6 Jul 2012 21:26:36 +0000 (UTC) Subject: Proposed update to the OpenJDK Web Site Terms of Use References: <20120705225355.1BC45888@eggemoggin.niobe.net> Message-ID: <63653466.11989.1341609996262.JavaMail.mail@webmail05> On Jul 6, 2012, mark.reinhold at oracle.com wrote: > 2012/7/5 14:31 -0700, mark at klomp.org: >> I assume it is fine to post something >> referencing one or the other (e.g. posting something like "I tried >> implementing the FooBar property for all Baz classes as described in >> paragraph 42, I won't post the code on this list, but it can be seen >> over there in my public repo and is available under the GPL, evaluating >> the included benchmark shows that... etc.") > > Yes. IANAL, but I would think that's perfectly fine. OK, cumbersome, but that is what you get when the expert group tries to stay inside the JCP with one leg. So it is fine for anyone to create GPLed implementations from the export group discussions to evaluate the proposed standard, but the export group then doesn't want direct access to that implementation. >>> The Expert Groups for such JSRs >>> will hold their technical discussions in the open, for all to see. The >>> specifications they create will continue to be licensed under terms >>> similar to those used in the past [2], but no click-through agreement >>> will be required in order to access specification materials. >> >> I take this to mean that you don't need to agree to the terms that are >> incompatible with creating a GPL implementation anymore just to read >> the now public specifications. That is really, really useful. Thanks. > > You don't need to agree explicitly to the comment-and-evaluation terms by > clicking on an "I agree" button. They're still there, though, whether > you agree with them or not. Sure, they are there for those who wish to play inside the JCP regime. The terms are incompatible with the default GPL license of the OpenJDK project. But you don't need to accept them and they aren't needed for people just wanting to create implementations of the open public standard documents. Thanks, Mark From sebastian.sickelmann at gmx.de Sat Jul 21 09:34:53 2012 From: sebastian.sickelmann at gmx.de (Sebastian Sickelmann) Date: Sat, 21 Jul 2012 11:34:53 +0200 Subject: Remove public fields in public API in a binary compatible way. Message-ID: <500A77BD.1020208@gmx.de> Hi all, sometime ago (in oct 2011[0]) I suggested a way how we can change(restrict) the visibility of fields in a binary compatible way. I had worked on the concepts some time and created some documentation[1] and an small demo[2]. The documentation uses some parts of the JEP-Template. But it is not intented to be a JEP yet. I think it is in an much to early state for a real JEP (isn't it, you can answer it after you had read it). The main parts I want to discuss is how to handle Reflection (should it be supported?), ExceptionHandling and the pros and cons of the two solutions I tried to describe. Some words to the demo. It uses ByteCodeTransformation at load time to change the access of fields to an invokedynamic call. The invokedynamic bootstrap method figures out which Method to call instead of the field access. The demo(AllInOneTest) uses the classes TestOld, TestNewSolution2, TestNew2Solution12, TestNew3Solution12 which got all the same exact implementation(byte-code), just that TestOld is testing the class OLD, TestNewSolution2 is testing NEWSol2, ... and TestNew3Solution12 is testing NEW3Sol12 The testclasses are generated using the class GEN. The GEN class generates two static testmethods: genStaticTest: which produces the following testsequence: System.out.println(testAgainst.staticField); testAgainst.staticField = new Object; System.out.println(testAgainst.staticField); genNonStaticTest: which produces the following testsequence: testAgainst tObject = new testAgainst(); System.out.println(tObject.cause); tObject.cause = new RuntimeException("NEW"); System.out.println(tObject.cause); Some words to the Implementation classes OLD, NEWSol2, ... , NEW3Sol12. OLD is the original form which we want to change. NEWSol2 is an complete compatible version of OLD, just the annotations are missing. It is compatible to Solution 2 only because it uses the old fieldnames (cause,staticField) NEW2Sol12 is an binary compatible version of OLD, just the annotations are missing. It is compatible to Solution 1 and 2 because it uses the old fieldnames (cause,staticField) NEW2Sol12 throws an runtime exception on every PUT of field "cause" and every PUTSTATIC of field "staticField" NEW3Sol12 is an binary and reflection compatible version of OLD. It is compatible to Solution 1 and 2 because it uses the old fieldnames (cause,staticField) NEW2Sol12 throws an checked exception on every PUT of field "cause" and every PUTSTATIC of field "staticField" Q0: What is this the right mailing-list to discuss this? Q1: What do you think about the main idea of this? Q2: So what do you think about Solution 1 and 2. Where are your pros and cons? Q3: Do you think we need to support reflection compabibility? Q4: What should we do if the accessor-method throws an CheckedException? Wrap it inside another RuntimeException(ex .WrappedCheckedCompatiblityException)? Q5: Should we also wrap RuntimeExceptions or should we let those "fly-through"? Hope to get some good discussion. -- Sebastian [0] http://mail.openjdk.java.net/pipermail/jdk8-dev/2011-October/000199.html [1] https://github.com/picpromusic/incubator/blob/master/jdk/compatibleFieldAccess/JEP.markdown [2] https://github.com/picpromusic/incubator/tree/master/jdk/compatibleFieldAccess From neil.richards at ngmr.net Wed Jul 25 11:15:04 2012 From: neil.richards at ngmr.net (Neil Richards) Date: Wed, 25 Jul 2012 12:15:04 +0100 Subject: Mailing lists for specification of enhanced metadata in Java SE 8 In-Reply-To: <500F0134.3070706@oracle.com> References: <50070F01.9080300@oracle.com> <5009AD07.1000000@oracle.com> <1343146995.18970.21.camel@chalkhill> <500F0134.3070706@oracle.com> Message-ID: <1343214904.18970.189.camel@chalkhill> Hi Alex, Thanks for your reply. It really helped to clarify how the two-list (-experts & -observers) approach operates. By observation, it seems that most activities in OpenJDK have used the (unified) -discuss model for their mailing lists, which is why, I guess, I'm unfamiliar with this alternative. In general, I haven't noticed conversations in these forums spinning wildly out of control due to rogue input from the ill-informed, but as I'm not claiming to be an expert, perhaps my spidey-sense is not attuned to it. Perhaps experts for other language-related lists such as lambda-dev or mlvm-dev might have a view on whether running discussion on a single list is unduly painful in this respect ? The few exceptions to this one-list way of doing things (in OpenJDK) seem to be those mailing lists directly related to specific JSRs. There the division between experts and observers is clearer, I suppose: experts are those on the Expert Group of the JSR, observers everyone else. (i.e. for JSRs, the mailing list structure just reflects that coming from the JSR itself). But as you're looking for this activity to be covered by the Java 8 Umbrella JSR (337), I'm not sure this makes sense here, as I don't think you're suggesting the membership on -experts is that of 337's EG. (In any case, 337 already has its own mailing lists). Also, in a two-list approach, I guess I'd still hope that the experts at least follow the traffic in the -observers list. Otherwise, there's not much point to discussing things there as there'd be scant chance of any good input in -observers getting adopted by the experts. Having said all that, I very much appreciate the notion of holding the design / spec conversations on a separate list to those for its implementation in OpenJDK, so thanks for looking to cater for that. Please don't interpret what I'm saying as active hostility to what you propose. I suppose it's just that I suspect you may not gain that much from the overhead it introduces. Incidentally, will the list of experts for this work map to some OpenJDK artifact, e.g. membership of an OpenJDK Project, or something ? Will it be clear how new people might apply / be proposed / join these experts ? (The OpenJDK Bylaws lay out the mechanics of how this is done for OpenJDK artifacts, so it's all clear how this would work if there is such a mapping). Regards, Neil On Tue, 2012-07-24 at 13:10 -0700, Alex Buckley wrote: > Hi Neil, > > On 7/24/2012 9:23 AM, Neil Richards wrote: > > Hi Alex, > > > > Once created, will these mailing lists appear (for subscription) on > > mail.openjdk.java.net, alongside the other (existing) lists ? > > Yes. > > The only reason there are two lists is to stop people from subscribing > to -experts and diverting the experts' discussion. This is a non-trivial > problem wherever Java language design is involved. I realize it's not > the friendliest thing in the world. > > > Is your intent that (the traffic on) both mailing lists are publicly > > archived, or is -experts traffic to be more "closed". > > Both lists are publicly archived. There are no secrets on -experts. > -observers is simply the mechanism by which -experts is followed if you > want an email record. If you want to read on the Web, just visit the > -experts archive. > > > Would you expect the experts to subscribe to both lists, so they can > > see/respond to discussion in -observers ? > > There is no expectation that members of -experts follow -observers. This > is analogous to the pair of lists for the SE 8 Expert Group. > > > Would you expect interested observers to subscribe to both lists, so > > they can see the pronouncements of the experts in -experts ? > > Interested parties can't subscribe to -experts; that's the point. If > they want an email record, they subscribe to -observers, which receives > a copy of everything on -experts. > > > Do you expect the traffic on -experts to be moderated, whilst that on > > -observers is unmoderated ?' > > Yes. > > > I'm trying to understand the working distinction between these two > > lists, so I understand (as an interested observer) how I should interact > > with them. > > You effectively cannot interact with -experts. In email, you can only > see its traffic via -observers, which you can comment too. > > Since sending out my proposal, I have been wondering whether it's > simpler to forget all about it and instead have a single > enhanced-metadata-spec-discuss list. Anyone could join and read and post > and leave as they wished. There would be no moderation. Posters would > still need to sign the OCA before making material contributions. I don't > know whether drive-by contributions from non-OCA signatories on a list > populated by mainly OCA signatories will raise legal issues. Comments? > > Alex -- Unless stated above: IBM email: neil_richards at uk.ibm.com IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU From amolk112k at gmail.com Fri Jul 27 11:20:10 2012 From: amolk112k at gmail.com (Amol Kulkarni) Date: Fri, 27 Jul 2012 16:50:10 +0530 Subject: =?GB2312?B?UmU6ILTwuLQ6ILTwuLQ6IDEwMCUgY3B1IHVzYWdlIG9uIENlbnRPUy9SSEVMIDY=?= In-Reply-To: <6DECD2A517016948873E76418A593E7009A02A0B@EXCHANGE-DAG2.xiaomi.net> References: <4FF1A9E6.4040503@optrak.com> <6DECD2A517016948873E76418A593E7009A026C9@EXCHANGE-DAG2.xiaomi.net> <6DECD2A517016948873E76418A593E7009A02A0B@EXCHANGE-DAG2.xiaomi.net> Message-ID: Hi, The problem was solved after upgrading CentOS kernel to kernel-2.6.32-220. This is the kernel that comes by default in CentOS 6.2. Thanks everyone for all the help. Amol On Tue, Jul 3, 2012 at 7:35 PM, ?? wrote: > If so, it's unlikely a leap-second issue. You can figure it out by > `dmesg`, the leap-second worked while sth like "Clock: inserting leap > second 23:59:60 UTC" be found and ntpd service running:) > ------------------------------ > *???:* Amol Kulkarni [amolk112k at gmail.com] > *????:* 2012?7?3? 21:54 > *?:* ?? > *Cc:* Mark Thornton; discuss at openjdk.java.net > *??:* Re: ??: 100% cpu usage on CentOS/RHEL 6 > > Seems like I've hit the leap second bug itself. But since I'm getting > that error for atleast 15 days now i.e from 15 June or so. Is it likely > that the leap second bug occurs before the leap second is introduced by the > kernel? > > On Tue, Jul 3, 2012 at 4:54 PM, ?? wrote: > >> May be you can have an eye on here : https://lkml.org/lkml/2012/7/1/27 :) >> ________________________________________ >> ???: discuss-bounces at openjdk.java.net [discuss-bounces at openjdk.java.net] >> ?? Amol Kulkarni [amolk112k at gmail.com] >> ????: 2012?7?3? 19:02 >> ?: Mark Thornton >> Cc: discuss at openjdk.java.net >> ??: Re: 100% cpu usage on CentOS/RHEL 6 >> >> Hi, >> >> >From the symptoms, it looks like the same bug to me too. >> I'll check whether it is solved in CentOS/RHEL 6 kernels. >> >> >> >> >> On Mon, Jul 2, 2012 at 7:32 PM, Mark Thornton >> wrote: >> >> > On 02/07/12 14:45, Amol Kulkarni wrote: >> > >> > Hi, >> > >> > I have a java app running on CentOS 6.0. It always runs in background >> via >> > cron. Sometimes this app goes into wait state while using 100% cpu. >> > >> > My java version is : >> > >> > java version "1.6.0_17" >> > OpenJDK Runtime Environment (IcedTea6 1.7.4) >> (rhel-1.21.b17.el6-x86_64) >> > OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode) >> > >> > Other symptoms are : >> > >> > a. One thread of the process seems to be in a loop waiting for >> something. >> > When traced using strace, it shows following o/p continuously : >> > >> > >> > futex(0x7fb8000ac728, FUTEX_WAKE_PRIVATE, 1) = 0 >> > futex(0x7fb8000ac754, >> FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, >> > >> > >> > Could it be related to this: >> > https://bugzilla.kernel.org/show_bug.cgi?id=32922 >> > >> > Mark Thornton >> > >> > >> > >