Re: Running in cluster mode causes native library linking to fail
Hello guys,
After lots of time trying to make things work, I finally found what was
causing the issue:
I was calling the function from the library inside a map function, which
caused the code inside it to be run in executors instead of the driver.
Since only the driver had loaded the library, the executors would then give
an error. The tricky part is that the same error message was being
replicated in the driver's and the executor's logs, so it led me to believe
it was a global error. Only after testing running stuff *only* on the
driver was that I discovered that everything worked.
For future reference: if you are running into this issue, please check if
you are also loading the library on the executors! In the case of my map
example, the fix was to create a wrapper function that 1) loaded libraries
and then 2) called functions within the library. After that, map things to
this wrapper function. This way, you ensure every executor also loads the
libraries.
I'd like to thank Prajod, Renato and Deenar for the help.
Bernardo
On 15 October 2015 at 03:27, <prajod.vettiyat...@wipro.com> wrote:
> Forwarding to the group, in case someone else has the same error. Just
> found out that I did not reply to the group in my original reply.
>
>
>
> *From:* Prajod S Vettiyattil (WT01 - BAS)
> *Sent:* 15 October 2015 11:45
> *To:* 'Bernardo Vecchia Stein' <bernardovst...@gmail.com>
> *Subject:* RE: Running in cluster mode causes native library linking to
> fail
>
>
>
> Hi,
>
>
>
> Also try the path settings given here:
> http://stackoverflow.com/questions/12279833/videocapture-opencv-2-4-2-error-in-windows/29278322#29278322
>
>
>
> Forgot to add this link in my response earlier:
>
> https://blogs.oracle.com/darcy/entry/purging_ld_library_path
>
> http://www.oracle.com/technetwork/java/javase/jdk7-relnotes-418459.html
>
>
>
> So from java 7, LD_LIBRARY_PATH is ignored. This is for Linux and Solaris.
> And probably for all other Unix derivatives.
>
>
>
> Also check : System.loadLibrary() should be inside a static { } block.
> Please check for its syntax on the internet. The loadlibrary function has
> to be called during class load time. That is why the static block is
> required.
>
>
>
> What is your ?
>
> 1. Spark version
>
> 2. OS type and version
>
> 3. Library that you are trying to load.
>
>
>
> [I was using OpenCV. Had to go through many trials to get it working
> consistently. Initially, it would work only on dev environment(windows) but
> not on Ubuntu. Its been a few months. There is a stackoverflow answer I
> have given regarding this:
> http://stackoverflow.com/questions/12279833/videocapture-opencv-2-4-2-error-in-windows/29278322#29278322
> ]
>
>
>
> Regards,
>
> Prajod
>
>
>
> *From:* Bernardo Vecchia Stein [mailto:bernardovst...@gmail.com]
> *Sent:* 15 October 2015 00:36
> *To:* Prajod S Vettiyattil (WT01 - BAS) <prajod.vettiyat...@wipro.com>
> *Subject:* Re: Running in cluster mode causes native library linking to
> fail
>
>
>
> Hello Prajod,
>
> Thanks for your reply! I am also using the standalone cluster manager. I
> do not build the jars in Eclipse and neither use Maven. They are built with
> sbt by hand.
>
> I was setting LD_LIBRARY_PATH and LIBRARY_PATH to point to the paths with
> the library. When I didn't set them and set only PATH instead, spark would
> just not find the libraries (it was another error). I'm not sure what
> version you are using, but it appears I do have to set LD_LIBRARY_PATH in
> order to make things work.
>
> I tried a simpler approach using System.load() with a specific path to the
> library, so I don't have to deal with these paths. However, I still get the
> same error when executing in cluster mode (exactly the same error). Do you
> have any idea why that might be failing?
>
> Thank you again for your attention,
>
> Bernardo
>
>
>
> On 14 October 2015 at 03:30, <prajod.vettiyat...@wipro.com> wrote:
>
> Hi,
>
>
>
> I have successfully made this working using the “standalone”cluster
> manager. Not tried with Mesos or YARN.
>
>
>
> Which of these cluster managers are you using ?
> https://spark.apache.org/docs/1.1.0/cluster-overview.html
>
> ·Standalone
> <https://spark.apache.org/docs/1.1.0/spark-standalone.html> – a simple
> cluster manager included with Spark that makes it easy to set up a cluster.
>
> ·Apache Mesos
> <https://spark.apache.org/docs/1.1.0/running-on-mesos.html> – a general
> cluster manager that can also run Hadoop MapReduce and service applications.
>
> ·Hadoop YARN
> <https://spark.apache.org/docs/1.1.0/r
Re: Running in cluster mode causes native library linking to fail
Hi Deenar, Yes, the native library is installed on all machines of the cluster. I tried a simpler approach by just using System.load() and passing the exact path of the library, and things still won't work (I get exactly the same error and message). Any ideas of what might be failing? Thank you, Bernardo On 14 October 2015 at 02:50, Deenar Toraskar <deenar.toras...@gmail.com> wrote: > Hi Bernardo > > Is the native library installed on all machines of your cluster and are > you setting both the spark.driver.extraLibraryPath and > spark.executor.extraLibraryPath ? > > Deenar > > > > On 14 October 2015 at 05:44, Bernardo Vecchia Stein < > bernardovst...@gmail.com> wrote: > >> Hello, >> >> I am trying to run some scala code in cluster mode using spark-submit. >> This code uses addLibrary to link with a .so that exists in the machine, >> and this library has a function to be called natively (there's a native >> definition as needed in the code). >> >> The problem I'm facing is: whenever I try to run this code in cluster >> mode, spark fails with the following message when trying to execute the >> native function: >> java.lang.UnsatisfiedLinkError: >> org.name.othername.ClassName.nativeMethod([B[B)[B >> >> Apparently, the library is being found by spark, but the required >> function isn't found. >> >> When trying to run in client mode, however, this doesn't fail and >> everything works as expected. >> >> Does anybody have any idea of what might be the problem here? Is there >> any bug that could be related to this when running in cluster mode? >> >> I appreciate any help. >> Thanks, >> Bernardo >> > >
Re: Running in cluster mode causes native library linking to fail
Hi Renato, I am using a single master and a single worker node, both in the same machine, to simplify everything. I have tested with System.loadLibrary() as well (setting all the necessary paths) and get the same error. Just double checked everything and the parameters are fine. Bernardo On 14 October 2015 at 16:54, Renato Marroquín Mogrovejo < renatoj.marroq...@gmail.com> wrote: > Sorry Bernardo, I just double checked. I use: System.loadLibrary(); > Could you also try that? > > > Renato M. > > 2015-10-14 21:51 GMT+02:00 Renato Marroquín Mogrovejo < > renatoj.marroq...@gmail.com>: > >> Hi Bernardo, >> >> So is this in distributed mode? or single node? Maybe fix the issue with >> a single node first ;) >> You are right that Spark finds the library but not the *.so file. I also >> use System.load() with LD_LIBRARY_PATH set, and I am able to >> execute without issues. Maybe you'd like to double check paths, env >> variables, or the parameters spark.driver.extraLibraryPath, >> spark.executor.extraLibraryPath. >> >> >> Best, >> >> Renato M. >> >> 2015-10-14 21:40 GMT+02:00 Bernardo Vecchia Stein < >> bernardovst...@gmail.com>: >> >>> Hi Renato, >>> >>> I have done that as well, but so far no luck. I believe spark is finding >>> the library correctly, otherwise the error message would be "no libraryname >>> found" or something like that. The problem seems to be something else, and >>> I'm not sure how to find it. >>> >>> Thanks, >>> Bernardo >>> >>> On 14 October 2015 at 16:28, Renato Marroquín Mogrovejo < >>> renatoj.marroq...@gmail.com> wrote: >>> >>>> You can also try setting the env variable LD_LIBRARY_PATH to point >>>> where your compiled libraries are. >>>> >>>> >>>> Renato M. >>>> >>>> 2015-10-14 21:07 GMT+02:00 Bernardo Vecchia Stein < >>>> bernardovst...@gmail.com>: >>>> >>>>> Hi Deenar, >>>>> >>>>> Yes, the native library is installed on all machines of the cluster. I >>>>> tried a simpler approach by just using System.load() and passing the exact >>>>> path of the library, and things still won't work (I get exactly the same >>>>> error and message). >>>>> >>>>> Any ideas of what might be failing? >>>>> >>>>> Thank you, >>>>> Bernardo >>>>> >>>>> On 14 October 2015 at 02:50, Deenar Toraskar < >>>>> deenar.toras...@gmail.com> wrote: >>>>> >>>>>> Hi Bernardo >>>>>> >>>>>> Is the native library installed on all machines of your cluster and >>>>>> are you setting both the spark.driver.extraLibraryPath and >>>>>> spark.executor.extraLibraryPath ? >>>>>> >>>>>> Deenar >>>>>> >>>>>> >>>>>> >>>>>> On 14 October 2015 at 05:44, Bernardo Vecchia Stein < >>>>>> bernardovst...@gmail.com> wrote: >>>>>> >>>>>>> Hello, >>>>>>> >>>>>>> I am trying to run some scala code in cluster mode using >>>>>>> spark-submit. This code uses addLibrary to link with a .so that exists >>>>>>> in >>>>>>> the machine, and this library has a function to be called natively >>>>>>> (there's >>>>>>> a native definition as needed in the code). >>>>>>> >>>>>>> The problem I'm facing is: whenever I try to run this code in >>>>>>> cluster mode, spark fails with the following message when trying to >>>>>>> execute >>>>>>> the native function: >>>>>>> java.lang.UnsatisfiedLinkError: >>>>>>> org.name.othername.ClassName.nativeMethod([B[B)[B >>>>>>> >>>>>>> Apparently, the library is being found by spark, but the required >>>>>>> function isn't found. >>>>>>> >>>>>>> When trying to run in client mode, however, this doesn't fail and >>>>>>> everything works as expected. >>>>>>> >>>>>>> Does anybody have any idea of what might be the problem here? Is >>>>>>> there any bug that could be related to this when running in cluster >>>>>>> mode? >>>>>>> >>>>>>> I appreciate any help. >>>>>>> Thanks, >>>>>>> Bernardo >>>>>>> >>>>>> >>>>>> >>>>> >>>> >>> >> >
Re: Running in cluster mode causes native library linking to fail
Hi Renato, I have done that as well, but so far no luck. I believe spark is finding the library correctly, otherwise the error message would be "no libraryname found" or something like that. The problem seems to be something else, and I'm not sure how to find it. Thanks, Bernardo On 14 October 2015 at 16:28, Renato Marroquín Mogrovejo < renatoj.marroq...@gmail.com> wrote: > You can also try setting the env variable LD_LIBRARY_PATH to point where > your compiled libraries are. > > > Renato M. > > 2015-10-14 21:07 GMT+02:00 Bernardo Vecchia Stein < > bernardovst...@gmail.com>: > >> Hi Deenar, >> >> Yes, the native library is installed on all machines of the cluster. I >> tried a simpler approach by just using System.load() and passing the exact >> path of the library, and things still won't work (I get exactly the same >> error and message). >> >> Any ideas of what might be failing? >> >> Thank you, >> Bernardo >> >> On 14 October 2015 at 02:50, Deenar Toraskar <deenar.toras...@gmail.com> >> wrote: >> >>> Hi Bernardo >>> >>> Is the native library installed on all machines of your cluster and are >>> you setting both the spark.driver.extraLibraryPath and >>> spark.executor.extraLibraryPath ? >>> >>> Deenar >>> >>> >>> >>> On 14 October 2015 at 05:44, Bernardo Vecchia Stein < >>> bernardovst...@gmail.com> wrote: >>> >>>> Hello, >>>> >>>> I am trying to run some scala code in cluster mode using spark-submit. >>>> This code uses addLibrary to link with a .so that exists in the machine, >>>> and this library has a function to be called natively (there's a native >>>> definition as needed in the code). >>>> >>>> The problem I'm facing is: whenever I try to run this code in cluster >>>> mode, spark fails with the following message when trying to execute the >>>> native function: >>>> java.lang.UnsatisfiedLinkError: >>>> org.name.othername.ClassName.nativeMethod([B[B)[B >>>> >>>> Apparently, the library is being found by spark, but the required >>>> function isn't found. >>>> >>>> When trying to run in client mode, however, this doesn't fail and >>>> everything works as expected. >>>> >>>> Does anybody have any idea of what might be the problem here? Is there >>>> any bug that could be related to this when running in cluster mode? >>>> >>>> I appreciate any help. >>>> Thanks, >>>> Bernardo >>>> >>> >>> >> >
Running in cluster mode causes native library linking to fail
Hello, I am trying to run some scala code in cluster mode using spark-submit. This code uses addLibrary to link with a .so that exists in the machine, and this library has a function to be called natively (there's a native definition as needed in the code). The problem I'm facing is: whenever I try to run this code in cluster mode, spark fails with the following message when trying to execute the native function: java.lang.UnsatisfiedLinkError: org.name.othername.ClassName.nativeMethod([B[B)[B Apparently, the library is being found by spark, but the required function isn't found. When trying to run in client mode, however, this doesn't fail and everything works as expected. Does anybody have any idea of what might be the problem here? Is there any bug that could be related to this when running in cluster mode? I appreciate any help. Thanks, Bernardo
Spark scala addFile retrieving file with incorrect size
Hi all, I'm trying to run a spark job (written in scala) that uses addFile to download some small files to each node. However, one of the downloaded files has an incorrect size (the other ones are ok), which causes an error when using it in the code. I have looked more into the issue and hexdump'ed both the original and the spark-retrieved files. The beginning of the files are exactly equal, but the spark-retrieved one just gets truncated at a random position. This position appears random, however I noticed that it is exactly half the size of the original file. Not sure if a coincidence or not. The original file has a size of 296 bytes (the others are a little bit bigger, around 13 kbytes). I'm kinda new to spark, so I'm stuck at this point trying to figure out what is the problem. Does anyone have any idea of what might be the problem here? Thank you, Bernardo
