CPU usage going upto 95%+ using Hazelcast

[namita nair]

Hi All,

         When i ran a test sample using hazelcast to cache data from DB i found that the CPU usage by java went up as high as 95% +. Can you please let me know is it a known issue.

Thanks and Regards,

Namita

[Enes Akar]

There was a problem on CPU usage but it has been fixed. 

Can you try your code with 2.6?

--
You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hazelcast+...@googlegroups.com.
To post to this group, send email to haze...@googlegroups.com.
Visit this group at http://groups.google.com/group/hazelcast.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--

Enes Akar

Hazelcast | Open source in-memory data grid
Mobile: +90.505.394.1668

[namita nair]

Hi All,

I am using hazelcast-2.6-ee currently.Its going high with this as well

Thanks and Regards,

Namita

[Enes Akar]

Have you enabled off-heap?

IF yes can you also try without off-heap?

--

You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hazelcast+...@googlegroups.com.
To post to this group, send email to haze...@googlegroups.com.
Visit this group at http://groups.google.com/group/hazelcast.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[namita nair]

Hi All,

My testing is with off heap.The elastic memory is been disabled.

Thanks and Regards,

Namita

On Wednesday, July 3, 2013 5:47:02 PM UTC+5:30, namita nair wrote:

[namita nair]

HI All,

We have decided to go with the heap as the test scenarios have proved to be better in case of the heap compared to the offheap. The responses were faster with heap.Is there any effect on the CPU usage if Heap is being used.Also we noticed when off heap is used the I/O wait goes high.Please suggest.

Thanks and Regards,

Namita

On Wednesday, July 3, 2013 5:47:02 PM UTC+5:30, namita nair wrote:

[Enes Akar]

Yes, using off-heap memory causes more CPU consumption.

If your memory per node is below 10 Gb, then we recommend you to use heap memory.

--

You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hazelcast+...@googlegroups.com.
To post to this group, send email to haze...@googlegroups.com.
Visit this group at http://groups.google.com/group/hazelcast.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[namita nair]

I used off heap as well 

Still the CPU usage show to be 95%+

On Wednesday, July 3, 2013 5:47:02 PM UTC+5:30, namita nair wrote:

[namita nair]

Hi All,

 

                Following is our machine details  :

 

Linux version : 2.6.32-131.4.1.el6.x86_64 (mock...@x86-003.build.bos.redhat.com) (gcc version 4.4.5 20110214 (Red Hat 4.4.5-6) (GCC) ) #1 SMP Fri Jun 10 10:54:26 EDT 2011

 

Top command output when we are trying to access data from the cache

 

top - 16:47:14 up 20 days,  3:39,  4 users,  load average: 0.84, 0.76, 0.56

Tasks: 160 total,   1 running, 159 sleeping,   0 stopped,   0 zombie

Cpu(s): 85.0%us,  5.3%sy,  0.0%ni,  8.3%id,  0.7%wa,  0.3%hi,  0.3%si,  0.0%st

Mem:   3288340k total,  2460780k used,   827560k free,     6020k buffers

Swap:  4193272k total,   194308k used,  3998964k free,   786192k cached

 

  PID USER      PR  NI  VIRT        RES       SHR S    %CPU  %MEM    TIME+       COMMAND

30022 root      20   0    1987m   597m   18m S     85.8      18.6          0:44.87      java

29905 root      20   0    2422m      704m   17m S   4.3        22.0         0:27.12      java

23804 root      20   0    1722m      173m  172m S  0.7       5.4            77:27.95   timestensubd

30358 root      20   0    15028       1220m    888 R  0.3       0.0            0:00.05      top

 

Please find the visualVM snapshots as an attachment.

Please let us know what can be done to reduce the CPU %.Any changes needed please let us know.

 

Thanks and Regards,

Namita

On Wednesday, July 3, 2013 5:47:02 PM UTC+5:30, namita nair wrote:

[namita nair]

Hi All,

           Any help on this would be apprciated. Since its blocking us.

Thanks and Regards,

Namita

On Wednesday, July 3, 2013 5:47:02 PM UTC+5:30, namita nair wrote:

[Enes Akar]

Do you have a simplified test code to simulate the behaviour?

--

You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hazelcast+...@googlegroups.com.
To post to this group, send email to haze...@googlegroups.com.
Visit this group at http://groups.google.com/group/hazelcast.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--

Enes Akar

Hazelcast | Open source in-memory data grid

Mobile: +90.507.150.56.71

[namita nair]

I solved the issue by making the following changes in the default.xml

 <property name="hazelcast.executor.client.thread.count">40</property>

 <property name="hazelcast.executor.event.thread.count">1</property>

 <property name="hazelcast.executor.migration.thread.count">1</property>

 <property name="hazelcast.executor.query.thread.count">1</property>

 <property name="hazelcast.executor.store.thread.count">1</property>

  <property name="hazelcast.heartbeat.interval.seconds">3600</property>

Please let me know if its correct and each parameter means its my assumption i made the changes.

I could not understand the meaning to each in the documentation.Are there and  other parameter that can changed to improve the performance?