This is an archive of the Maximo Yahoo Community. The content of this pages may be a sometimes obsolete so please check post dates.
Thanks to the community owner Christopher Wanko for providing the content.
Our JVMs are configured to use 4GB of ram and has the following arguments
-Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
--- In MAXIMO@yahoogroups.com, Yahoo wrote:
>
> What are the java heap settings in websphere?
>
>
>
> Shannon Sutton
> Cell: (318) 334-8002
>
> On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
>
> > Fellow Maximo users,
> >
> > The City of Bellevue is working to improve Maximo 7.5 system performance and would like to know your experience with tuning Maximo. Here is what our users are experiencing:
> > User sessions timeout
> > Slow to return results(user query)
> > Slow to tab between fields
> > Slow to show value lists
> > Start centers slow to load
> > Long running processes fail to refresh screen
> > Hour glass/freezes random times
> > Running large/complex queries affect other users
> >
> > Our current environment runs on Windows 2008 R2 servers. We run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs running on the same physical server (2 quad core 32GB ram). On the same server we have one JVM dedicated to inbound transactions and Cron tasks. Print jobs are handled by a JVM running on a seperate physical server. All sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the desktop.
> >
> > We would appreciate any input â" and we are happy to reciprocate.
> >
> > Jason Hale
> > IT Analyst
> > City of Bellevue
> >
> >
>
>
>
>
How many concurrent users do you have at peak load? Are they evenly distributed across the JVMs?
It's sounding as though either the db is struggling or the app server is swapping to disk.
Chris.
On Jan 26, 2013, at 7:53 PM, "jhale1996" <jhale1996@yahoo.com> wrote:
> Our JVMs are configured to use 4GB of ram and has the following arguments
>
> -Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
>
> --- In MAXIMO@yahoogroups.com, Yahoo wrote:
> >
> > What are the java heap settings in websphere?
> >
> >
> >
> > Shannon Sutton
> > Cell: (318) 334-8002
> >
> > On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
> >
> > > Fellow Maximo users,
> > >
> > > The City of Bellevue is working to improve Maximo 7.5 system performance and would like to know your experience with tuning Maximo. Here is what our users are experiencing:
> > > User sessions timeout
> > > Slow to return results(user query)
> > > Slow to tab between fields
> > > Slow to show value lists
> > > Start centers slow to load
> > > Long running processes fail to refresh screen
> > > Hour glass/freezes random times
> > > Running large/complex queries affect other users
> > >
> > > Our current environment runs on Windows 2008 R2 servers. We run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs running on the same physical server (2 quad core 32GB ram). On the same server we have one JVM dedicated to inbound transactions and Cron tasks. Print jobs are handled by a JVM running on a seperate physical server. All sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the desktop.
> > >
> > > We would appreciate any input â€" and we are happy to reciprocate.
> > >
> > > Jason Hale
> > > IT Analyst
> > > City of Bellevue
> > >
> > >
> >
> >
> >
> >
>
>
At peak load we are close to 80 users. Distribution is fairly uniform.
--- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
>
> How many concurrent users do you have at peak load? Are they evenly distributed across the JVMs?
>
> It's sounding as though either the db is struggling or the app server is swapping to disk.
>
> Chris.
>
> On Jan 26, 2013, at 7:53 PM, "jhale1996" wrote:
>
> > Our JVMs are configured to use 4GB of ram and has the following arguments
> >
> > -Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
> >
> > --- In MAXIMO@yahoogroups.com, Yahoo wrote:
> > >
> > > What are the java heap settings in websphere?
> > >
> > >
> > >
> > > Shannon Sutton
> > > Cell: (318) 334-8002
> > >
> > > On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
> > >
> > > > Fellow Maximo users,
> > > >
> > > > The City of Bellevue is working to improve Maximo 7.5 system performance and would like to know your experience with tuning Maximo. Here is what our users are experiencing:
> > > > User sessions timeout
> > > > Slow to return results(user query)
> > > > Slow to tab between fields
> > > > Slow to show value lists
> > > > Start centers slow to load
> > > > Long running processes fail to refresh screen
> > > > Hour glass/freezes random times
> > > > Running large/complex queries affect other users
> > > >
> > > > Our current environment runs on Windows 2008 R2 servers. We run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs running on the same physical server (2 quad core 32GB ram). On the same server we have one JVM dedicated to inbound transactions and Cron tasks. Print jobs are handled by a JVM running on a seperate physical server. All sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the desktop.
> > > >
> > > > We would appreciate any input ââ¬" and we are happy to reciprocate.
> > > >
> > > > Jason Hale
> > > > IT Analyst
> > > > City of Bellevue
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
>
>
>
>
You should be fine with two or three UI JVMs and still have contingency. Download the IBM Support Assistant and the GC analyzer or, if you prefer, zip and email me your native_stderr.log files and I'll chart them for you.
On Jan 26, 2013, at 8:32 PM, "jhale1996" <jhale1996@yahoo.com> wrote:
> At peak load we are close to 80 users. Distribution is fairly uniform.
>
> --- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
> >
> > How many concurrent users do you have at peak load? Are they evenly distributed across the JVMs?
> >
> > It's sounding as though either the db is struggling or the app server is swapping to disk.
> >
> > Chris.
> >
> > On Jan 26, 2013, at 7:53 PM, "jhale1996" wrote:
> >
> > > Our JVMs are configured to use 4GB of ram and has the following arguments
> > >
> > > -Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
> > >
> > > --- In MAXIMO@yahoogroups.com, Yahoo wrote:
> > > >
> > > > What are the java heap settings in websphere?
> > > >
> > > >
> > > >
> > > > Shannon Sutton
> > > > Cell: (318) 334-8002
> > > >
> > > > On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
> > > >
> > > > > Fellow Maximo users,
> > > > >
> > > > > The City of Bellevue is working to improve Maximo 7.5 system performance and would like to know your experience with tuning Maximo. Here is what our users are experiencing:
> > > > > User sessions timeout
> > > > > Slow to return results(user query)
> > > > > Slow to tab between fields
> > > > > Slow to show value lists
> > > > > Start centers slow to load
> > > > > Long running processes fail to refresh screen
> > > > > Hour glass/freezes random times
> > > > > Running large/complex queries affect other users
> > > > >
> > > > > Our current environment runs on Windows 2008 R2 servers. We run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs running on the same physical server (2 quad core 32GB ram). On the same server we have one JVM dedicated to inbound transactions and Cron tasks. Print jobs are handled by a JVM running on a seperate physical server. All sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the desktop.
> > > > >
> > > > > We would appreciate any input â€" and we are happy to reciprocate.
> > > > >
> > > > > Jason Hale
> > > > > IT Analyst
> > > > > City of Bellevue
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> >
> >
> >
> >
>
>
Thanks for the information on the GC analyzer. I've downloaded the GC analyzer and ran the logs through it I get a similiar result on each node:
File name : \\cobappslts28\d$\ibm\WebSphere\AppServer\profiles\ctgAppSrv01\logs\MAXIMOPRODUI5\native_stderr.log
Number of verboseGC cycles : 6
Number of Garbage Collections : 4416
Number of Allocation failures : 4412
First Garbage Collection : Mon Dec 17 19:37:19 2012
Last Garbage Collection : Mon Jan 28 07:51:23 2013
Number of Java heap exhaustion : 0
Overall Garbage Collection overhead : 0.02%
Maximum Garbage Collection overhead : 100% (Thu Dec 20 07:30:51 2012)
Number of 100% AF overhead : 4
Total Garbage Collection pause : 808 seconds
Maximum Tenured Area usage : 3,089,785,824 bytes (Fri Jan 11 14:01:06 2013)
Average Tenured Area usage : 1,715,829,651 bytes
Number of Explicit Garbage Collection : 0
Maximum Allocation Request : 11,256,696 bytes (Thu Jan 17 17:57:17 2013)
Number of Object Requests larger than 10 M bytes : 1
Java Heap Activity Analysis and Recommendations report
Garbage collection start / finish
Analysis
Recommendations
#1 Mon Dec 17 19:37:19 2012 Thu Dec 20 16:08:41 2012 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 31.669703 with percentage error(%): 0.33754677)
#2 Thu Dec 20 16:16:14 2012 Thu Dec 20 18:57:01 2012 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 70.1474 with percentage error(%): 3.599103)
#3 Thu Dec 20 19:50:24 2012 Sun Jan 13 18:28:33 2013 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 216.1234 with percentage error(%): 0.38971308)
#4 Sun Jan 13 19:41:31 2013 Sun Jan 13 19:42:35 2013 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 11.559267 with percentage error(%): 2.4091938)
#5 Sun Jan 13 19:59:03 2013 Wed Jan 16 18:01:05 2013 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 993.82825 with percentage error(%): 0.46781746)
#6 Wed Jan 16 21:32:21 2013 Mon Jan 28 07:51:23 2013 Configuration gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes initialHeapSize : 4,294,967,296 bytes compressedRefs : true compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096 requestedPageSize : 4,096
No Java heap exhaustion found
There seems to be a steady increase in Java heap usage. ( ratio(%): 728.34076 with percentage error(%): 0.37917668)
--- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
>
> You should be fine with two or three UI JVMs and still have contingency. Download the IBM Support Assistant and the GC analyzer or, if you prefer, zip and email me your native_stderr.log files and I'll chart them for you.
>
> On Jan 26, 2013, at 8:32 PM, "jhale1996" wrote:
>
> > At peak load we are close to 80 users. Distribution is fairly uniform.
> >
> > --- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
> > >
> > > How many concurrent users do you have at peak load? Are they evenly distributed across the JVMs?
> > >
> > > It's sounding as though either the db is struggling or the app server is swapping to disk.
> > >
> > > Chris.
> > >
> > > On Jan 26, 2013, at 7:53 PM, "jhale1996" wrote:
> > >
> > > > Our JVMs are configured to use 4GB of ram and has the following arguments
> > > >
> > > > -Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
> > > >
> > > > --- In MAXIMO@yahoogroups.com, Yahoo wrote:
> > > > >
> > > > > What are the java heap settings in websphere?
> > > > >
> > > > >
> > > > >
> > > > > Shannon Sutton
> > > > > Cell: (318) 334-8002
> > > > >
> > > > > On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
> > > > >
> > > > > > Fellow Maximo users,
> > > > > >
> > > > > > The City of Bellevue is working to improve Maximo 7.5 system performance and would like to know your experience with tuning Maximo. Here is what our users are experiencing:
> > > > > > User sessions timeout
> > > > > > Slow to return results(user query)
> > > > > > Slow to tab between fields
> > > > > > Slow to show value lists
> > > > > > Start centers slow to load
> > > > > > Long running processes fail to refresh screen
> > > > > > Hour glass/freezes random times
> > > > > > Running large/complex queries affect other users
> > > > > >
> > > > > > Our current environment runs on Windows 2008 R2 servers. We run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs running on the same physical server (2 quad core 32GB ram). On the same server we have one JVM dedicated to inbound transactions and Cron tasks. Print jobs are handled by a JVM running on a seperate physical server. All sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the desktop.
> > > > > >
> > > > > > We would appreciate any input ââ¬" and we are happy to reciprocate.
> > > > > >
> > > > > > Jason Hale
> > > > > > IT Analyst
> > > > > > City of Bellevue
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
>
>
>
>
How high does "available heap after collection" get? Anywhere near 4Gb?
On Mon, Jan 28, 2013 at 10:59 AM, jhale1996 <jhale1996@yahoo.com> wrote:
> **
>
>
> Thanks for the information on the GC analyzer. I've downloaded the GC
> analyzer and ran the logs through it I get a similiar result on each node:
>
> File name :
> \\cobappslts28\d$\ibm\WebSphere\AppServer\profiles\ctgAppSrv01\logs\MAXIMOPRODUI5\native_stderr.log
> Number of verboseGC cycles : 6
> Number of Garbage Collections : 4416
> Number of Allocation failures : 4412
> First Garbage Collection : Mon Dec 17 19:37:19 2012
> Last Garbage Collection : Mon Jan 28 07:51:23 2013
> Number of Java heap exhaustion : 0
> Overall Garbage Collection overhead : 0.02%
> Maximum Garbage Collection overhead : 100% (Thu Dec 20 07:30:51 2012)
> Number of 100% AF overhead : 4
> Total Garbage Collection pause : 808 seconds
> Maximum Tenured Area usage : 3,089,785,824 bytes (Fri Jan 11 14:01:06 2013)
> Average Tenured Area usage : 1,715,829,651 bytes
> Number of Explicit Garbage Collection : 0
> Maximum Allocation Request : 11,256,696 bytes (Thu Jan 17 17:57:17 2013)
> Number of Object Requests larger than 10 M bytes : 1
>
> Java Heap Activity Analysis and Recommendations report
> Garbage collection start / finish
> Analysis
> Recommendations
> #1 Mon Dec 17 19:37:19 2012 Thu Dec 20 16:08:41 2012 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 31.669703 with percentage error(%): 0.33754677)
> #2 Thu Dec 20 16:16:14 2012 Thu Dec 20 18:57:01 2012 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 70.1474 with percentage error(%): 3.599103)
> #3 Thu Dec 20 19:50:24 2012 Sun Jan 13 18:28:33 2013 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 216.1234 with percentage error(%): 0.38971308)
> #4 Sun Jan 13 19:41:31 2013 Sun Jan 13 19:42:35 2013 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 11.559267 with percentage error(%): 2.4091938)
> #5 Sun Jan 13 19:59:03 2013 Wed Jan 16 18:01:05 2013 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 993.82825 with percentage error(%): 0.46781746)
> #6 Wed Jan 16 21:32:21 2013 Mon Jan 28 07:51:23 2013 Configuration
> gcPolicy : -Xgcpolicy:gencon maxHeapSize : 4,294,967,296 bytes
> initialHeapSize : 4,294,967,296 bytes compressedRefs : true
> compressedRefsDisplacement : 0 compressedRefsShift : 3 pageSize : 4,096
> requestedPageSize : 4,096
> No Java heap exhaustion found
> There seems to be a steady increase in Java heap usage. ( ratio(%):
> 728.34076 with percentage error(%): 0.37917668)
>
>
> --- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
> >
> > You should be fine with two or three UI JVMs and still have contingency.
> Download the IBM Support Assistant and the GC analyzer or, if you prefer,
> zip and email me your native_stderr.log files and I'll chart them for you.
> >
> > On Jan 26, 2013, at 8:32 PM, "jhale1996" wrote:
> >
> > > At peak load we are close to 80 users. Distribution is fairly uniform.
> > >
> > > --- In MAXIMO@yahoogroups.com, Chris Lawless wrote:
> > > >
> > > > How many concurrent users do you have at peak load? Are they evenly
> distributed across the JVMs?
> > > >
> > > > It's sounding as though either the db is struggling or the app
> server is swapping to disk.
> > > >
> > > > Chris.
> > > >
> > > > On Jan 26, 2013, at 7:53 PM, "jhale1996" wrote:
> > > >
> > > > > Our JVMs are configured to use 4GB of ram and has the following
> arguments
> > > > >
> > > > > -Dsun.rmi.dgc.ackTimeout=10000 -Djava.net.preferIPv4Stack=true
> -Xdisableexplicitgc -Xgcpolicy:gencon -Xmn1024m -verbose:gc
> > > > >
> > > > > --- In MAXIMO@yahoogroups.com, Yahoo wrote:
> > > > > >
> > > > > > What are the java heap settings in websphere?
> > > > > >
> > > > > >
> > > > > >
> > > > > > Shannon Sutton
> > > > > > Cell: (318) 334-8002
> > > > > >
> > > > > > On Jan 24, 2013, at 10:58 PM, "jhale1996" wrote:
> > > > > >
> > > > > > > Fellow Maximo users,
> > > > > > >
> > > > > > > The City of Bellevue is working to improve Maximo 7.5 system
> performance and would like to know your experience with tuning Maximo. Here
> is what our users are experiencing:
> > > > > > > User sessions timeout
> > > > > > > Slow to return results(user query)
> > > > > > > Slow to tab between fields
> > > > > > > Slow to show value lists
> > > > > > > Start centers slow to load
> > > > > > > Long running processes fail to refresh screen
> > > > > > > Hour glass/freezes random times
> > > > > > > Running large/complex queries affect other users
> > > > > > >
> > > > > > > Our current environment runs on Windows 2008 R2 servers. We
> run WebSphere 7.0.0.15 and MS SQL 2008. Our users connect to one of 5 JVMs
> running on the same physical server (2 quad core 32GB ram). On the same
> server we have one JVM dedicated to inbound transactions and Cron tasks.
> Print jobs are handled by a JVM running on a seperate physical server. All
> sites are connected with fiber, 1Gb to the servers and 100Mb cooper to the
> desktop.
> > > > > > >
> > > > > > > We would appreciate any input â€" and we are happy to
> reciprocate.
>
> > > > > > >
> > > > > > > Jason Hale
> > > > > > > IT Analyst
> > > > > > > City of Bellevue
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> >
> >
> >
> >
>
>
>