perfSONAR User Q&A and Other Discussion

[Aaron Brown <>]

Hi,

That’s presumably it. By default, cassandra will set the heap size to 1/4 of 
system memory, or 4G in your case. Since you’ve got a 32-bit host, a single 
process can’t allocate the 4G space, so it bombs out.

Cheers,
Aaron

On Oct 23, 2014, at 12:43 AM, Rod James Bio 
<>
 wrote:

> A work around that worked for me is to set the vars under 
> /etc/cassandra/default.conf/cassandra-env.sh to:
> 
> MAX_HEAP_SIZE="2048M"
> HEAP_NEWSIZE="400M"
> 
> I am using an 8core machine with 16GB RAM (12GB Free) BUT a 32-bit 
> perfsonar version. So that may be the issue (?).
> 
> Regards,
> 
> On 10/23/14, 12:27, Rod James Bio wrote:
>> Hello everyone,
>> 
>> I updated my perfsonar to 3.4 and it was returning an undefined after
>> checking the graphs, similar to this:
>> 
>> http://www.perfsonar.net/about/faq/#Q68
>> 
>> After doing the suggestion, i.e. restart cassandra and httpd, I noticed
>> that cassandra was not starting with this on the cassandra logfile:
>> 
>> Error occurred during initialization of VM
>> Could not reserve enough space for object heap
>> Error: Could not create the Java Virtual Machine.
>> Error: A fatal exception has occurred. Program will exit.
>> 
>> Does anyone have any suggestion on how to fix/work-around this?
>> 
>> Regards,