You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@kafka.apache.org by Vishnu Viswanath <vi...@gmail.com> on 2018/08/22 15:13:31 UTC
Re: RocksDB error with KStream
Thank you Guoshang. I finally solved the issue by shifting from openjdk to
https://hub.docker.com/r/anapsix/alpine-java/ in my container.
Looks like alpine openjdk has musl libc implementation but RocksJava
required glibc.
On Mon, Jul 30, 2018 at 9:31 AM Guozhang Wang <wa...@gmail.com> wrote:
> Hello Vishnu,
>
> Maybe this is the root cause of your issue?
> https://github.com/docker-flink/docker-flink/issues/14
>
>
> Guozhang
>
> On Mon, Jul 30, 2018 at 8:44 AM, Vishnu Viswanath <
> vishnu.viswanath25@gmail.com> wrote:
>
> > Hi All,
> >
> > While running Kafka streams application with a "aggregate" function in
> it.
> > Running via docker (alpine linux). I am getting
> > *"*java.lang.UnsatisfiedLinkError:
> > /tmp/librocksdbjni20844899025353663*88.so: libstdc++.so.6: cannot open
> > shared object file: No such file or directory". *similar to what is
> > described here:
> >
> > https://groups.google.com/forum/#!msg/confluent-platform/uFTFZh3v-kg/
> > luNsCcRbBQAJ
> >
> >
> > In the Dockerfile I did create a temp directory with +w permission and
> pass
> > it as "-Djava.io.tmpdir". If I login into the instance, I could see that
> > librocksdbjni2084489902535366388.so file is created in the tmpdir, but
> > application still fails with unsatisfied link error.
> >
> > Could solve it by changing to in memory store, but is skeptical if this
> > will be an issue later if the aggregate function uses lot of memory. Is
> > there any other suggestion on solving this problem (i.e., use rocksdb as
> > state store and run within a docker container)
> >
> > Thanks in advance,
> > Vishnu
> >
>
>
>
> --
> -- Guozhang
>
Re: RocksDB error with KStream
Posted by Guozhang Wang <wa...@gmail.com>.
Glad to hear that!
On Wed, Aug 22, 2018 at 8:13 AM, Vishnu Viswanath <
vishnu.viswanath25@gmail.com> wrote:
> Thank you Guoshang. I finally solved the issue by shifting from openjdk to
> https://hub.docker.com/r/anapsix/alpine-java/ in my container.
> Looks like alpine openjdk has musl libc implementation but RocksJava
> required glibc.
>
> On Mon, Jul 30, 2018 at 9:31 AM Guozhang Wang <wa...@gmail.com> wrote:
>
> > Hello Vishnu,
> >
> > Maybe this is the root cause of your issue?
> > https://github.com/docker-flink/docker-flink/issues/14
> >
> >
> > Guozhang
> >
> > On Mon, Jul 30, 2018 at 8:44 AM, Vishnu Viswanath <
> > vishnu.viswanath25@gmail.com> wrote:
> >
> > > Hi All,
> > >
> > > While running Kafka streams application with a "aggregate" function in
> > it.
> > > Running via docker (alpine linux). I am getting
> > > *"*java.lang.UnsatisfiedLinkError:
> > > /tmp/librocksdbjni20844899025353663*88.so: libstdc++.so.6: cannot open
> > > shared object file: No such file or directory". *similar to what is
> > > described here:
> > >
> > > https://groups.google.com/forum/#!msg/confluent-platform/uFTFZh3v-kg/
> > > luNsCcRbBQAJ
> > >
> > >
> > > In the Dockerfile I did create a temp directory with +w permission and
> > pass
> > > it as "-Djava.io.tmpdir". If I login into the instance, I could see
> that
> > > librocksdbjni2084489902535366388.so file is created in the tmpdir, but
> > > application still fails with unsatisfied link error.
> > >
> > > Could solve it by changing to in memory store, but is skeptical if this
> > > will be an issue later if the aggregate function uses lot of memory. Is
> > > there any other suggestion on solving this problem (i.e., use rocksdb
> as
> > > state store and run within a docker container)
> > >
> > > Thanks in advance,
> > > Vishnu
> > >
> >
> >
> >
> > --
> > -- Guozhang
> >
>
--
-- Guozhang