You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@cassandra.apache.org by Yang <te...@gmail.com> on 2011/08/05 04:38:57 UTC

what is taking up my JVM heap space?

I ran my cassandra cluster on one node, inserted about 500000 rows,
each about 400 bytes, then the heap went up to 800MB, and constantly
goes into full GC. I guessed that the memtable was probably taking a
lot of space, so I did a node tool flush, so that
all memtable are changed to 0. after this, the heap came down a bit,  to 500MB.

but I thought it should be close to 0 right after the flush?? or else
what is taking up the space?
I'm still trying to look at the hprof file to figure out, but it seems
difficult too


Thanks
Yang

Re: what is taking up my JVM heap space?

Posted by Yang <te...@gmail.com>.

Thanks Jonathan


On Thu, Aug 4, 2011 at 7:44 PM, Jonathan Ellis <jb...@gmail.com> wrote:
> If you didn't GC post-flush then it's just leftovers.
>
> An "empty" Cassandra (nothing in memtables or cache) will use 5-10MB of heap.
>
> On Thu, Aug 4, 2011 at 9:38 PM, Yang <te...@gmail.com> wrote:
>> I ran my cassandra cluster on one node, inserted about 500000 rows,
>> each about 400 bytes, then the heap went up to 800MB, and constantly
>> goes into full GC. I guessed that the memtable was probably taking a
>> lot of space, so I did a node tool flush, so that
>> all memtable are changed to 0. after this, the heap came down a bit,  to 500MB.
>>
>> but I thought it should be close to 0 right after the flush?? or else
>> what is taking up the space?
>> I'm still trying to look at the hprof file to figure out, but it seems
>> difficult too
>>
>>
>> Thanks
>> Yang
>>
>
>
>
> --
> Jonathan Ellis
> Project Chair, Apache Cassandra
> co-founder of DataStax, the source for professional Cassandra support
> http://www.datastax.com
>

Re: what is taking up my JVM heap space?

Posted by Jonathan Ellis <jb...@gmail.com>.

If you didn't GC post-flush then it's just leftovers.

An "empty" Cassandra (nothing in memtables or cache) will use 5-10MB of heap.

On Thu, Aug 4, 2011 at 9:38 PM, Yang <te...@gmail.com> wrote:
> I ran my cassandra cluster on one node, inserted about 500000 rows,
> each about 400 bytes, then the heap went up to 800MB, and constantly
> goes into full GC. I guessed that the memtable was probably taking a
> lot of space, so I did a node tool flush, so that
> all memtable are changed to 0. after this, the heap came down a bit,  to 500MB.
>
> but I thought it should be close to 0 right after the flush?? or else
> what is taking up the space?
> I'm still trying to look at the hprof file to figure out, but it seems
> difficult too
>
>
> Thanks
> Yang
>



-- 
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of DataStax, the source for professional Cassandra support
http://www.datastax.com