HI Ganesh,
Recommendation is to increase the number of nodes with lesser ram size.
Your number of executors depend on the CPU core hence, i would recommend
using 60 GB RAM cpu's with 2 executors each for your use case.
Thanks
Ramesh
On Wed, Feb 22, 2017 at 3:27 PM, Sai Ganesh Muthuraman <
saigan
Hi Team,
I have been trying out of core graph option with 1.2 branch of giraph . I
have been
continously recieving the buffer underflow exception exactly when the
superstep1 starts. Any pointers would be really helpful here ,
Logs:
2016-07-25 16:08:20,664 ERROR [ooc-io-0]
org.apache.giraph.utils.
Hi,
This does not seem to be a dependecy issue . This is an exception in launch
of the container. Can you check the yarn configurations.
Thanks,
Ramesh
On Sun, May 22, 2016 at 9:52 PM, Rahman, Muntasir Raihan <
mrahm...@illinois.edu> wrote:
> Thanks for the suggestion, but unfortunately it did
Hi ,
Container Launch exception happens usually because of either dependency
missing issues or an actual error while launching a container. The best
way to know the error accurately is to look at the YARN logs very
precisely. Can you put the last 200 lines of the yarn logs for the
corresponding
pushed to the code base yet. You
> can access this feature here: https://reviews.facebook.net/D55479
>
> Best,
> Hassan
>
> On Sat, May 14, 2016 at 10:46 PM, Ramesh Krishnan > wrote:
>
>> Thanks Hassan. I have removed the checkpointing, still getting a
>> different
eckpointing yet. This is part of an ongoing project, and we should have
> the integration within a few weeks. In the meantime, you can try
> out-of-core without checkpointing enabled.
>
> Best,
> Hassan
>
>
> On Saturday, May 14, 2016, Ramesh Krishnan
> wrote:
>
>>
)
at java.lang.Thread.run(Thread.java:745)
On Sun, May 15, 2016 at 1:02 AM, Ramesh Krishnan
wrote:
>
> Hi Team,
>
> I have the latest build of giraph running on a 5 node cluster. When i try
> to use OutofCore Graph option for a huge data set like 600Milion edges i am
Hi Team,
I have the latest build of giraph running on a 5 node cluster. When i try
to use OutofCore Graph option for a huge data set like 600Milion edges i am
running into
the following exception. Please find below the script being executed and
the exception logs. I have tried all possible ways a