Thanks. Trying it on my local machine now but it will probably take a
while. I think https://github.com/apache/spark/pull/34085 is more likely to
be relevant but don't yet have a clue how it could cause the issue. Spark
CI also passed for these.
On Mon, Sep 27, 2021 at 5:29 PM Sean Owen wrote:
I'm building and testing with
mvn -Phadoop-3.2 -Phive -Phive-2.3 -Phive-thriftserver -Pkinesis-asl
-Pkubernetes -Pmesos -Pnetlib-lgpl -Pscala-2.12 -Pspark-ganglia-lgpl
-Psparkr -Pyarn ...
I did a '-DskipTests clean install' and then 'test'; the problem arises
only in 'test'.
On Mon, Sep 27,
Hmm it may be related to the commit. Sean: how do I reproduce this?
On Mon, Sep 27, 2021 at 4:56 PM Sean Owen wrote:
> Another "is anyone else seeing this"? in compiling common/yarn-network:
>
> [ERROR] [Error]
>
Another "is anyone else seeing this"? in compiling common/yarn-network:
[ERROR] [Error]
/mnt/data/testing/spark-3.2.0/common/network-yarn/src/main/java/org/apache/spark/network/yarn/YarnShuffleService.java:32:
package com.google.common.annotations does not exist
[ERROR] [Error]
I think even if we do cancel this RC we should leave it open for a bit to
see if we can catch any other errors.
On Mon, Sep 27, 2021 at 12:29 PM Dongjoon Hyun
wrote:
> Unfortunately, it's the same for me recently. Not only that, but I also
> hit MetaspaceSize OOM, too.
> I ended up with
Unfortunately, it's the same for me recently. Not only that, but I also hit
MetaspaceSize OOM, too.
I ended up with MAVEN_OPTS like the following.
-Xms12g -Xmx12g -Xss128M -XX:MaxMetaspaceSize=4g ...
Dongjoon.
On Mon, Sep 27, 2021 at 12:18 PM Sean Owen wrote:
> Has anyone seen a
Has anyone seen a StackOverflowError when running tests? It happens in
compilation. I heard from another user who hit this earlier, and I had not,
until just today testing this:
[ERROR] ## Exception when compiling 495 sources to
/mnt/data/testing/spark-3.2.0/sql/catalyst/target/scala-2.12/classes
Hi Kousuke,
I tend to agree with Sean. It only affects the macOS developers when
building Spark with the released Spark 3.2 code tarball without setting
JAVA_HOME.
I can mention this one as a known issue in the release note if this vote
passes.
Thanks,
Gengliang
On Mon, Sep 27, 2021 at 11:47 PM
I think it affects devs but there are some workarounds.
So, if you all don't think it's necessary to meet 3.2.0, I'm OK not to
do it.
- Kousuke
Hm... it does just affect Mac OS (?) and only if you don't have
JAVA_HOME set (which people often do set) and only affects build/mvn,
vs built-in
Hm... it does just affect Mac OS (?) and only if you don't have JAVA_HOME
set (which people often do set) and only affects build/mvn, vs built-in
maven (which people often have installed). Only affects those building. I'm
on the fence about whether it blocks 3.2.0, as it doesn't affect downstream
Hi All,
SPARK-35887 seems to have introduced another issue that building with
build/mvn on macOS stucks, and SPARK-36856 will resolve this issue.
Should we meet the fix to 3.2.0?
- Kousuke
Please vote on releasing the following candidate as Apache Spark
version 3.2.0.
The vote is open
Starting with my +1(non-binding)
Thanks,
Gengliang
On Mon, Sep 27, 2021 at 8:55 PM Gengliang Wang wrote:
> Please vote on releasing the following candidate as
> Apache Spark version 3.2.0.
>
> The vote is open until 11:59pm Pacific time September 29 and passes if a
> majority +1 PMC votes are
Please vote on releasing the following candidate as
Apache Spark version 3.2.0.
The vote is open until 11:59pm Pacific time September 29 and passes if a
majority +1 PMC votes are cast, with a minimum of 3 +1 votes.
[ ] +1 Release this package as Apache Spark 3.2.0
[ ] -1 Do not release this
13 matches
Mail list logo
