Liangcai Li
Liangcai Li
> Curious, why the stack showed it called GPU cudf code, but the repro code contains spark.conf.set("spark.rapids.sql.enabled", false) to disable the GPU? Is it correct? @mythrocks It is enabled just...
> ~Ah, right. I should have enabled the plugin. Thanks for catching it. I'll correct the description.~ Edit: Nope, the script is correct, as it is. > > I'm trying...
I tried locally (non-db env) and it passed with Spark3.3.0 + delta-2.1.0 + plugin-22.10. But not sure if it actually ran into the path that causes the crash. ``` ____...
Thanks ttnghia, I will try it on DB.
Found the root cause. See https://github.com/rapidsai/cudf/issues/11700. I am trying to make a fix ...
Here is the fix https://github.com/rapidsai/cudf/pull/11702
> > I tried locally (non-db env) and it passed with Spark3.3.0 + delta-2.1.0 + plugin-22.10. > > Sorry I'm late to this party, but I can confirm that this...
> > The stacktrace that Mithun posted above should be somewhat relevant to the partitioning issue. > > Ah, thanks! I didn't look far enough back in the issue, missed...
> @firestarman do you have the stacktrace where we're trying to partition an empty table? There's checks in the plugin where we have avoided trying to shuffle empty tables, curious...
Hi, I saw the latest version is updating the doc only, does it mean this needs to be fixed in the JNI ?