Hi Julien,
Thanks for the suggestion. If we don't do a broadcast, that would
presumably affect the performance of the job, as the model that is failing
to be broadcast is something that we need to be shared across the cluster.
But it may be worth it if the trade-off is not having things run proper
Hi,
Did you try without the broadcast ?
Regards
JL
Le jeu. 19 sept. 2019 à 06:41, Vadim Semenov
a écrit :
> Pre-register your classes:
>
> ```
> import com.esotericsoftware.kryo.Kryo
> import org.apache.spark.serializer.KryoRegistrator
>
> class MyKryoRegistrator extends KryoRegistrator {
> ov
I remember it not working for us when we were setting it from the inside
and needed to actually pass it
On Wed, Sep 18, 2019 at 10:38 AM Jerry Vinokurov
wrote:
> Hi Vadim,
>
> Thanks for your suggestion. We do preregister the classes, like so:
>
> object KryoRegistrar {
>>
>> val classesToRegi
Hi Vadim,
Thanks for your suggestion. We do preregister the classes, like so:
object KryoRegistrar {
>
> val classesToRegister: Array[Class[_]] = Array(
> classOf[MyModel],
>[etc]
> ) }
>
And then we do:
val sparkConf = new SparkConf()
> .registerKryoClasses(KryoRegistrar.classe
Pre-register your classes:
```
import com.esotericsoftware.kryo.Kryo
import org.apache.spark.serializer.KryoRegistrator
class MyKryoRegistrator extends KryoRegistrator {
override def registerClasses(kryo: Kryo): Unit = {
kryo.register(Class.forName("[[B")) // byte[][]
kryo.register(clas
Hi folks,
Posted this some time ago but the problem continues to bedevil us. I'm
including a (slightly edited) stack trace that results from this error. If
anyone can shed any light on what exactly is happening here and what we can
do to avoid it, that would be much appreciated.
org.apache.spark.
