From sushantkumar76 at gmail.com Thu Sep 3 14:26:27 2015 From: sushantkumar76 at gmail.com (susantra m) Date: Thu, 3 Sep 2015 19:56:27 +0530 Subject: G1 PermGen Usage for JAVA 7 Message-ID: Hello Everyone, I have a question related to how G1 handles permgen till JAVA7 . Basically till JAVA7 ,permgen was part of the JAVA heap , so as the heap is divided into regions in G1 where exactly is permgen is stored and how G1 manages that . It seems G1 divides heap into regions , so if i have 2GB of Heap which is divided into regions , i would assume Permgen to be part of the regions till i am using JDK7 . With metaspace it is clear cut that it is outside of heap . Though i am assuming it is maintaining permgen outside of the heap and they are not part of the regions , need some more info on this . Need some more clarification on this and Appreciate any help and suggestions. Thanks, Sushant. -------------- next part -------------- An HTML attachment was scrubbed... URL: From poonam.bajaj at oracle.com Thu Sep 3 17:51:54 2015 From: poonam.bajaj at oracle.com (Poonam Bajaj Parhar) Date: Thu, 03 Sep 2015 10:51:54 -0700 Subject: G1 PermGen Usage for JAVA 7 In-Reply-To: References: Message-ID: <55E888BA.3080602@oracle.com> Hello Sushant, If we run with -XX:+PrintGCDetails option, it shows the GC algorithm used for all the generations. D:\Java\jdk1.7.0_40\bin>java.exe -XX:+UseG1GC -XX:+PrintGCDetails -version java version "1.7.0_40" Java(TM) SE Runtime Environment (build 1.7.0_40-b43) Java HotSpot(TM) Client VM (build 24.0-b56, mixed mode) Heap garbage-first heap total 16384K, used 0K [0x04d00000, 0x05d00000, 0x14d00000) region size 1024K, 1 young (1024K), 0 survivors (0K) compacting perm gen total 12288K, used 1354K [0x14d00000, 0x15900000, 0x18d00000) the space 12288K, 11% used [0x14d00000, 0x14e52a88, 0x14e52c00, 0x15900000) No shared spaces configured. Here we can see that PermGen is out of G1 heap, and G1GC algorithm is not used for the Permanent generation. Thanks, Poonam On 9/3/2015 7:26 AM, susantra m wrote: > > Hello Everyone, > I have a question related to how G1 handles permgen till JAVA7 . > Basically till JAVA7 ,permgen was part of the JAVA heap , so as the > heap is divided into regions in G1 where exactly is permgen is stored > and how G1 manages that . > > It seems G1 divides heap into regions , so if i have 2GB of Heap > which is divided into regions , i would assume Permgen to be part of > the regions till i am using JDK7 . With metaspace it is clear cut > that it is outside of heap . > > Though i am assuming it is maintaining permgen outside of the heap > and they are not part of the regions , need some more info on this . > > Need some more clarification on this and Appreciate any help and > suggestions. > > > Thanks, > Sushant. > > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use -------------- next part -------------- An HTML attachment was scrubbed... URL: From naman.jain1894 at yahoo.in Tue Sep 8 08:40:45 2015 From: naman.jain1894 at yahoo.in (Naman Jain) Date: Tue, 8 Sep 2015 08:40:45 +0000 (UTC) Subject: Reclaim of dead humongous objects during young GC. Message-ID: <718952879.2865135.1441701645323.JavaMail.yahoo@mail.yahoo.com> Hey everyone, I just come across these two options G1ReclaimDeadHumongousObjectsAtYoungGC and G1TraceReclaimDeadHumongousObjectsAtYoungGC while looking for a solution to overcome humongous allocation in G1.Can some one tell me whether these options work in java version 1.8.0_40 and above or not.If yes, then does humongous object are collected during every young GC ? Naman? -------------- next part -------------- An HTML attachment was scrubbed... URL: From thomas.schatzl at oracle.com Tue Sep 8 09:20:00 2015 From: thomas.schatzl at oracle.com (Thomas Schatzl) Date: Tue, 08 Sep 2015 11:20:00 +0200 Subject: Reclaim of dead humongous objects during young GC. In-Reply-To: <718952879.2865135.1441701645323.JavaMail.yahoo@mail.yahoo.com> References: <718952879.2865135.1441701645323.JavaMail.yahoo@mail.yahoo.com> Message-ID: <1441704000.2262.12.camel@oracle.com> Hi, On Tue, 2015-09-08 at 08:40 +0000, Naman Jain wrote: > Hey everyone, > > I just come across these two options > G1ReclaimDeadHumongousObjectsAtYoungGC and > G1TraceReclaimDeadHumongousObjectsAtYoungGC while looking for a > solution to overcome humongous allocation in G1. > Can some one tell me whether these options work in java version > 1.8.0_40 and above or not. Yes - I even think the feature is enabled by default in 8u40. However we fixed some cases and improved it slightly in 8u60, so I would recommend using 8u60. > If yes, then does humongous object are collected during every young > GC ? G1 uses some heuristics to decide whether it should try to reclaim these regions. First, the technique only applies to arrays of primitive types (integer, bools, ...) at this time, and they should never have been referenced by many objects during its lifetime. It works best (actually extremely well) with applications that are transaction oriented, where these large temporary objects are kept local by a thread for things like output buffers. G1 always applies this technique at every young GC. Note that in 8u60, the switches have been renamed (iirc), i.e. G1EagerReclaimHumongousObjects to enable/disable use (default: true), G1EagerReclaimHumongousObjectsWithStaleRefs to enable some optimization for it to try to reclaim harder (default: true) and G1TraceEagerReclaimHumongousObjects (default: false). G1TraceEagerReclaimHumongousObjects gives some information why (and why not) particular large objects were reclaimed (or not). However already with maximum logging (-XX:G1LogLevel=finest) there is some indication about how effective the technique is/has been applied. Thanks, Thomas From josephj at arssoftware.com Thu Sep 17 13:02:38 2015 From: josephj at arssoftware.com (Joseph) Date: Thu, 17 Sep 2015 18:32:38 +0530 Subject: GC Tuning Message-ID: <075701d0f149$214793b0$63d6bb10$@com> Hi , I am Joseph. I am trying to tune performance of our server .I run server with following g1 gc options. Maximum heap allocated to server is 5GB. " -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps" The gc logs resulted with above gc flags are attached with the mail.How to reduce garbage pause time in our server. Thank you for your valuble time Regards, Joseph ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: gc-hbase.log Type: application/octet-stream Size: 16573 bytes Desc: not available URL: From poonam.bajaj at oracle.com Thu Sep 17 18:39:18 2015 From: poonam.bajaj at oracle.com (Poonam Bajaj Parhar) Date: Thu, 17 Sep 2015 11:39:18 -0700 Subject: GC Tuning In-Reply-To: <075701d0f149$214793b0$63d6bb10$@com> References: <075701d0f149$214793b0$63d6bb10$@com> Message-ID: <55FB08D6.9020206@oracle.com> Hello Joseph, From the GC logs, all the G1 evacuation pauses are under 100ms of pause time goal. However, there are System.gc() invoked Full GCs that are taking longer time. /62.730: [GC pause (young), 0.02510300 secs] ... 64.287: [Full GC (System.gc()) 175M->98M(2048M), 0.2183040 secs] [Times: user=0.26 sys=0.01, real=0.22 secs] 64.509: [Full GC (System.gc()) 98M->98M(2048M), 0.1928990 secs] [Times: user=0.24 sys=0.00, real=0.20 secs] 64.704: [Full GC (System.gc()) 98M->98M(2048M), 0.1893760 secs] [Times: user=0.25 sys=0.00, real=0.19 secs] 64.896: [Full GC (System.gc()) 98M->97M(2048M), 0.2001590 secs] [Times: user=0.25 sys=0.00, real=0.20 secs] 65.098: [Full GC (System.gc()) 98M->97M(2048M), 0.1874640 secs] [Times: user=0.24 sys=0.00, real=0.19 secs] 66.818: [Full GC (System.gc()) 185M->99M(2048M), 0.2137830 secs] [Times: user=0.27 sys=0.00, real=0.22 secs]// 73.240: [GC pause (young), 0.02516600 secs]/ Please check your application code to see if these System.gc() invocations are intentional. You may also disable these using -XX:+DisableExplicitGC option. Thanks, Poonam On 9/17/2015 6:02 AM, Joseph wrote: > > Hi , > > I am Joseph. I am trying to tune performance of our server .I run > server with following g1 gc options. > > Maximum heap allocated to server is 5GB. > > " -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 > -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled > -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log > -XX:+PrintGCDetails -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps" > > The gc logs resulted with above gc flags are attached with the > mail.How to reduce garbage pause time in our server. > > Thank you for your valuble time > > *Regards,* > > *Joseph* > > > ------------------------------------------------------------------------ > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use -------------- next part -------------- An HTML attachment was scrubbed... URL: From yiyeguhu at gmail.com Thu Sep 17 20:50:44 2015 From: yiyeguhu at gmail.com (Tao Mao) Date: Thu, 17 Sep 2015 13:50:44 -0700 Subject: GC Tuning In-Reply-To: <075701d0f149$214793b0$63d6bb10$@com> References: <075701d0f149$214793b0$63d6bb10$@com> Message-ID: Hi, What's the goal here (latency or throughput)? What's the bottom line now? Thanks. Tao Mao On Thu, Sep 17, 2015 at 6:02 AM, Joseph wrote: > Hi , > > > > I am Joseph. I am trying to tune performance of our server .I run server > with following g1 gc options. > > > > Maximum heap allocated to server is 5GB. > > > > ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 > -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled > -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails > -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? > > > > The gc logs resulted with above gc flags are attached with the mail.How to > reduce garbage pause time in our server. > > > > Thank you for your valuble time > > > > > > *Regards,* > > *Joseph* > > > > ------------------------------ > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From yiyeguhu at gmail.com Sat Sep 19 08:22:47 2015 From: yiyeguhu at gmail.com (Tao Mao) Date: Sat, 19 Sep 2015 01:22:47 -0700 Subject: GC Tuning In-Reply-To: <002b01d0f202$231772a0$694657e0$@com> References: <075701d0f149$214793b0$63d6bb10$@com> <002b01d0f202$231772a0$694657e0$@com> Message-ID: Hi Joseph, What's the average response time under current settings? What typical use cases are you expecting from HBase (get to know some basic idea of your application)? How large is your data set? Thanks. Tao Mao On Fri, Sep 18, 2015 at 4:06 AM, Joseph wrote: > Hi Tao, > > > > Thank you for your quick response. > > > > I am trying to tune the No SQL database which require real time response. > Here in some cases server not responding due to the garbage pauses. Maximum > response pause allowed is below or equal to 100. When I set -XX:MaxGCPauseMillis=100 > flag, each collection pause exceeds at most of 100 milliseconds?. I want > reduce the total pause time maximum of 100 milliseconds . Because server > needs to respond within milliseconds. > > > > > > Regards > > Joseph > > > > *From:* Tao Mao [mailto:yiyeguhu at gmail.com] > *Sent:* Friday, September 18, 2015 2:21 AM > *To:* Joseph James > *Cc:* hotspot-gc-use > *Subject:* Re: GC Tuning > > > > Hi, > > > > What's the goal here (latency or throughput)? What's the bottom line now? > > > > Thanks. > > Tao Mao > > > > > > On Thu, Sep 17, 2015 at 6:02 AM, Joseph wrote: > > Hi , > > > > I am Joseph. I am trying to tune performance of our server .I run server > with following g1 gc options. > > > > Maximum heap allocated to server is 5GB. > > > > ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 > -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled > -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails > -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? > > > > The gc logs resulted with above gc flags are attached with the mail.How to > reduce garbage pause time in our server. > > > > Thank you for your valuble time > > > > > > *Regards,* > > *Joseph* > > > > > ------------------------------ > > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use > > > > ------------------------------ > This mail is sent from the ARS T&TT(India)Pvt Ltd. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From josephj at arssoftware.com Fri Sep 18 11:06:58 2015 From: josephj at arssoftware.com (Joseph) Date: Fri, 18 Sep 2015 16:36:58 +0530 Subject: GC Tuning In-Reply-To: References: <075701d0f149$214793b0$63d6bb10$@com> Message-ID: <002b01d0f202$231772a0$694657e0$@com> Hi Tao, Thank you for your quick response. I am trying to tune the No SQL database which require real time response. Here in some cases server not responding due to the garbage pauses. Maximum response pause allowed is below or equal to 100. When I set -XX:MaxGCPauseMillis=100 flag, each collection pause exceeds at most of 100 milliseconds?. I want reduce the total pause time maximum of 100 milliseconds . Because server needs to respond within milliseconds. Regards Joseph From: Tao Mao [mailto:yiyeguhu at gmail.com] Sent: Friday, September 18, 2015 2:21 AM To: Joseph James Cc: hotspot-gc-use Subject: Re: GC Tuning Hi, What's the goal here (latency or throughput)? What's the bottom line now? Thanks. Tao Mao On Thu, Sep 17, 2015 at 6:02 AM, Joseph > wrote: Hi , I am Joseph. I am trying to tune performance of our server .I run server with following g1 gc options. Maximum heap allocated to server is 5GB. ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? The gc logs resulted with above gc flags are attached with the mail.How to reduce garbage pause time in our server. Thank you for your valuble time Regards, Joseph ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. _______________________________________________ hotspot-gc-use mailing list hotspot-gc-use at openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. -------------- next part -------------- An HTML attachment was scrubbed... URL: From josephj at arssoftware.com Fri Sep 18 11:16:43 2015 From: josephj at arssoftware.com (Joseph) Date: Fri, 18 Sep 2015 16:46:43 +0530 Subject: GC Tuning In-Reply-To: <55FB08D6.9020206@oracle.com> References: <075701d0f149$214793b0$63d6bb10$@com> <55FB08D6.9020206@oracle.com> Message-ID: <003001d0f203$7fae7df0$7f0b79d0$@com> Hi Poonam, Thank you for your valuable time. Using the flags mentioned below ,I am trying to tune the nosql database HBase .Due to this reason I have no control over the source code . I will update the flags and back with the updated logs. Thank you Regards Joseph From: Poonam Bajaj Parhar [mailto:poonam.bajaj at oracle.com] Sent: Friday, September 18, 2015 12:09 AM To: Joseph James; hotspot-gc-use at openjdk.java.net Subject: Re: GC Tuning Hello Joseph, >From the GC logs, all the G1 evacuation pauses are under 100ms of pause time goal. However, there are System.gc() invoked Full GCs that are taking longer time. 62.730: [GC pause (young), 0.02510300 secs] ... 64.287: [Full GC (System.gc()) 175M->98M(2048M), 0.2183040 secs] [Times: user=0.26 sys=0.01, real=0.22 secs] 64.509: [Full GC (System.gc()) 98M->98M(2048M), 0.1928990 secs] [Times: user=0.24 sys=0.00, real=0.20 secs] 64.704: [Full GC (System.gc()) 98M->98M(2048M), 0.1893760 secs] [Times: user=0.25 sys=0.00, real=0.19 secs] 64.896: [Full GC (System.gc()) 98M->97M(2048M), 0.2001590 secs] [Times: user=0.25 sys=0.00, real=0.20 secs] 65.098: [Full GC (System.gc()) 98M->97M(2048M), 0.1874640 secs] [Times: user=0.24 sys=0.00, real=0.19 secs] 66.818: [Full GC (System.gc()) 185M->99M(2048M), 0.2137830 secs] [Times: user=0.27 sys=0.00, real=0.22 secs] 73.240: [GC pause (young), 0.02516600 secs] Please check your application code to see if these System.gc() invocations are intentional. You may also disable these using -XX:+DisableExplicitGC option. Thanks, Poonam On 9/17/2015 6:02 AM, Joseph wrote: Hi , I am Joseph. I am trying to tune performance of our server .I run server with following g1 gc options. Maximum heap allocated to server is 5GB. " -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps" The gc logs resulted with above gc flags are attached with the mail.How to reduce garbage pause time in our server. Thank you for your valuble time Regards, Joseph ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. _______________________________________________ hotspot-gc-use mailing list hotspot-gc-use at openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. -------------- next part -------------- An HTML attachment was scrubbed... URL: From ecki at zusammenkunft.net Mon Sep 21 16:07:47 2015 From: ecki at zusammenkunft.net (Bernd) Date: Mon, 21 Sep 2015 18:07:47 +0200 Subject: GC Tuning In-Reply-To: <002b01d0f202$231772a0$694657e0$@com> References: <075701d0f149$214793b0$63d6bb10$@com> <002b01d0f202$231772a0$694657e0$@com> Message-ID: Josef, the -XX:MaxGCPauseMillis=100 is only a hint to the GC to aim for those pause times. You will not get a gurantee (and typically 100ms is a limit you can only reach on highly specialized setups (small heap, low occupancy setings, etc). In fact it has been reported that aiming too low for this target can even hurt. Gruss Bernd 2015-09-18 13:06 GMT+02:00 Joseph : > Hi Tao, > > > > Thank you for your quick response. > > > > I am trying to tune the No SQL database which require real time response. > Here in some cases server not responding due to the garbage pauses. Maximum > response pause allowed is below or equal to 100. When I set > -XX:MaxGCPauseMillis=100 flag, each collection pause exceeds at most of > 100 milliseconds?. I want reduce the total pause time maximum of 100 > milliseconds . Because server needs to respond within milliseconds. > > > > > > Regards > > Joseph > > > > From: Tao Mao [mailto:yiyeguhu at gmail.com] > Sent: Friday, September 18, 2015 2:21 AM > To: Joseph James > Cc: hotspot-gc-use > Subject: Re: GC Tuning > > > > Hi, > > > > What's the goal here (latency or throughput)? What's the bottom line now? > > > > Thanks. > > Tao Mao > > > > > > On Thu, Sep 17, 2015 at 6:02 AM, Joseph wrote: > > Hi , > > > > I am Joseph. I am trying to tune performance of our server .I run server > with following g1 gc options. > > > > Maximum heap allocated to server is 5GB. > > > > ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 > -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled > -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails > -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? > > > > The gc logs resulted with above gc flags are attached with the mail.How to > reduce garbage pause time in our server. > > > > Thank you for your valuble time > > > > > > Regards, > > Joseph > > > > > > ________________________________ > > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use > > > > > ________________________________ > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use > From yiyeguhu at gmail.com Tue Sep 22 21:58:25 2015 From: yiyeguhu at gmail.com (Tao Mao) Date: Tue, 22 Sep 2015 14:58:25 -0700 Subject: GC Tuning In-Reply-To: <00bb01d0f534$4ac36bd0$e04a4370$@com> References: <075701d0f149$214793b0$63d6bb10$@com> <002b01d0f202$231772a0$694657e0$@com> <00bb01d0f534$4ac36bd0$e04a4370$@com> Message-ID: Hi Joseph, HBase seems to deal with a lot more data typically; perhaps, you'd want to have a look at other DB solutions (I can provide a list if you want). But since I don't know about your specific use case and/or context it's possible that's the only solution given your circumstance. Coming back to your GC question, as others pointed out, the log indicates SystemGC-triggered Full GCs, which can be disabled by setting -XX:+DisableExplicitGC. Otherwise, young gc doesn't look like a problem for the requirement (i.e., query response time under 1 sec, gc time under 100ms) as it occurs about every 5~20 sec and lasts <100 ms. Please report back the findings after you modify the gc flags. Thanks. Tao Mao On Tue, Sep 22, 2015 at 5:43 AM, Joseph wrote: > Hi Tao, > > > > Servers currently respond within 1.2 seconds .But this response may take > more time, when there is garbage related issue. Servers logs ?response is > too slow? which is usually occur due to the garbage pauses. > > > > We need to deliver data within one second .Data size is not so large which > is approximately 3 MB . > > > > *From:* Tao Mao [mailto:yiyeguhu at gmail.com] > *Sent:* Saturday, September 19, 2015 8:23 AM > > *To:* Joseph James > *Cc:* hotspot-gc-use > *Subject:* Re: GC Tuning > > > > Hi Joseph, > > > > What's the average response time under current settings? > > > > What typical use cases are you expecting from HBase (get to know some > basic idea of your application)? How large is your data set? > > > > Thanks. > > Tao Mao > > > > On Fri, Sep 18, 2015 at 4:06 AM, Joseph wrote: > > Hi Tao, > > > > Thank you for your quick response. > > > > I am trying to tune the No SQL database which require real time response. > Here in some cases server not responding due to the garbage pauses. Maximum > response pause allowed is below or equal to 100. When I set -XX:MaxGCPauseMillis=100 > flag, each collection pause exceeds at most of 100 milliseconds?. I want > reduce the total pause time maximum of 100 milliseconds . Because server > needs to respond within milliseconds. > > > > > > Regards > > Joseph > > > > *From:* Tao Mao [mailto:yiyeguhu at gmail.com] > *Sent:* Friday, September 18, 2015 2:21 AM > *To:* Joseph James > *Cc:* hotspot-gc-use > *Subject:* Re: GC Tuning > > > > Hi, > > > > What's the goal here (latency or throughput)? What's the bottom line now? > > > > Thanks. > > Tao Mao > > > > > > On Thu, Sep 17, 2015 at 6:02 AM, Joseph wrote: > > Hi , > > > > I am Joseph. I am trying to tune performance of our server .I run server > with following g1 gc options. > > > > Maximum heap allocated to server is 5GB. > > > > ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 > -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled > -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails > -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? > > > > The gc logs resulted with above gc flags are attached with the mail.How to > reduce garbage pause time in our server. > > > > Thank you for your valuble time > > > > > > *Regards,* > > *Joseph* > > > > > ------------------------------ > > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > > _______________________________________________ > hotspot-gc-use mailing list > hotspot-gc-use at openjdk.java.net > http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use > > > > > ------------------------------ > > This mail is sent from the ARS T&TT(India)Pvt Ltd. > > > > ------------------------------ > This mail is sent from the ARS T&TT(India)Pvt Ltd. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From josephj at arssoftware.com Tue Sep 22 12:43:33 2015 From: josephj at arssoftware.com (Joseph) Date: Tue, 22 Sep 2015 12:43:33 +0000 Subject: GC Tuning In-Reply-To: References: <075701d0f149$214793b0$63d6bb10$@com> <002b01d0f202$231772a0$694657e0$@com> Message-ID: <00bb01d0f534$4ac36bd0$e04a4370$@com> Hi Tao, Servers currently respond within 1.2 seconds .But this response may take more time, when there is garbage related issue. Servers logs ?response is too slow? which is usually occur due to the garbage pauses. We need to deliver data within one second .Data size is not so large which is approximately 3 MB . From: Tao Mao [mailto:yiyeguhu at gmail.com] Sent: Saturday, September 19, 2015 8:23 AM To: Joseph James Cc: hotspot-gc-use Subject: Re: GC Tuning Hi Joseph, What's the average response time under current settings? What typical use cases are you expecting from HBase (get to know some basic idea of your application)? How large is your data set? Thanks. Tao Mao On Fri, Sep 18, 2015 at 4:06 AM, Joseph > wrote: Hi Tao, Thank you for your quick response. I am trying to tune the No SQL database which require real time response. Here in some cases server not responding due to the garbage pauses. Maximum response pause allowed is below or equal to 100. When I set -XX:MaxGCPauseMillis=100 flag, each collection pause exceeds at most of 100 milliseconds?. I want reduce the total pause time maximum of 100 milliseconds . Because server needs to respond within milliseconds. Regards Joseph From: Tao Mao [mailto:yiyeguhu at gmail.com] Sent: Friday, September 18, 2015 2:21 AM To: Joseph James Cc: hotspot-gc-use Subject: Re: GC Tuning Hi, What's the goal here (latency or throughput)? What's the bottom line now? Thanks. Tao Mao On Thu, Sep 17, 2015 at 6:02 AM, Joseph > wrote: Hi , I am Joseph. I am trying to tune performance of our server .I run server with following g1 gc options. Maximum heap allocated to server is 5GB. ? -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:ParallelGCThreads=4 -XX:InitiatingHeapOccupancyPercent=65 -XX:+ParallelRefProcEnabled -XX:-ResizePLAB -Xloggc:/var/log/hbase/gc-hbase.log -XX:+PrintGCDetails -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCTimeStamps? The gc logs resulted with above gc flags are attached with the mail.How to reduce garbage pause time in our server. Thank you for your valuble time Regards, Joseph ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. _______________________________________________ hotspot-gc-use mailing list hotspot-gc-use at openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. ________________________________ This mail is sent from the ARS T&TT(India)Pvt Ltd. -------------- next part -------------- An HTML attachment was scrubbed... URL: