Oh, the container image itself is very large. Do you have some place I can upload it to?
On Monday, October 17, 2016 8:43:10 AM CEST Avinash Sridharan wrote:
> Also would be good to see the env variables (LD_LIBRARY_PATH) being seen by
> the container when you start it, just to make sure the env are getting set
> correctly.
>
> On Mon, Oct 17, 2016 at 8:42 AM, Avinash Sridharan <avin...@mesosphere.io>
>
> wrote:
> > Does look like the symlink exists:
> > /usr/lib/libtiff.so.5
> >
> > I am assuming you have checked the realpath exists as well (that's why was
> > asking for `ls -al`) ?
> >
> > Don't see that you have volume mounts that might obfuscate the path. Could
> > you create a JIRA and if possible point give access to your docker image
> > for us to try? Do mention the exact version of Mesos and the Distro you
> > are
> > trying to run this on.
> >
> > -Avinash
> >
> > On Mon, Oct 17, 2016 at 8:35 AM, Mark Hammons <
> >
> > mark.hamm...@inaf.cnrs-gif.fr> wrote:
> >> No, it's a regular linux log. I've reattached it.
> >>
> >> On Monday, October 17, 2016 8:32:07 AM CEST Avinash Sridharan wrote:
> >> > can't seem to open the attached logs, is it gzip?
> >> >
> >> > On Mon, Oct 17, 2016 at 8:14 AM, Mark Hammons <
> >>
> >> mark.hamm...@inaf.cnrs-gif.fr
> >>
> >> > > wrote:
> >> > >
> >> > > Adding LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu doesn't fix the
> >>
> >> error.
> >>
> >> > > On Monday, October 17, 2016 5:06:34 PM CEST Mark Hammons wrote:
> >> > > > As I've shown in the original logs, the symbolic link libtiff.so.5
> >>
> >> is in
> >>
> >> > > > /usr/ lib. I used LD_PRELOAD just to try to force
> >>
> >> /usr/lib/libtiff.so.5
> >>
> >> > > to
> >> > >
> >> > > > be loaded since it wasn't for some reason.
> >> > > >
> >> > > > On Monday, October 17, 2016 8:03:34 AM CEST Avinash Sridharan
wrote:
> >> > > > > Can you prepend your command with `ls -al /usr/lib` and check in
> >> > >
> >> > > stdout if
> >> > >
> >> > > > > you are seeing the shared object? By the way why are you using
> >> > >
> >> > > LD_PRELOAD
> >> > >
> >> > > > > instead of LD_LIBRARY_PATH. LD_PRELOAD will force the linker to
> >>
> >> load
> >>
> >> > > your
> >> > >
> >> > > > > library before any other libraries. It's usually used to override
> >> > >
> >> > > system
> >> > >
> >> > > > > libraries so found a bit odd that you are using it here?
> >> > > > >
> >> > > > > On Mon, Oct 17, 2016 at 7:54 AM, Mark Hammons
> >> > > > > <mark.hamm...@inaf.cnrs-gif.fr>
> >> > > > >
> >> > > > > > wrote:
> >> > > > > >
> >> > > > > > No, but even when I set LD_PRELOAD=/usr/lib/libtiff.so.5 for
> >>
> >> the
> >>
> >> > > process
> >> > >
> >> > > > > > environment it says it can't load /usr/lib/libtiff.so.5.
> >> > > > > >
> >> > > > > > On Monday, October 17, 2016 7:52:16 AM CEST Avinash Sridharan
> >>
> >> wrote:
> >> > > > > > > Are you setting the env in the dockerfile?
> >> > > > > > >
> >> > > > > > > On Mon, Oct 17, 2016 at 6:58 AM, Mark Hammons <
> >> > > > > >
> >> > > > > > mark.hamm...@inaf.cnrs-gif.fr
> >> > > > > >
> >> > > > > > > > wrote:
> >> > > > > > > >
> >> > > > > > > > Here's the code I define my executor to mesos with:
> >> > > > > > > > val iuwtURI = CommandInfo.URI.newBuilder().
> >> > >
> >> > > setValue("http://
> >> > >
> >> > > > > > ***/
> >> > > > > >
> >> > > > > > > > IUWT.tar.gz").setExtract(true).setCache(false).build()
> >> > > > > > > >
> >> > > > > > > > val iuwtjURI =
> >> > > > > > > > CommandInfo.URI.newBuilder().setValue("http://
> >> > > > > >
> >> > > > > > ***/
> >> > > > > >
> >> > > > > > > > iuwtExecutor-assembly-0.1-
> >> > > > > > > > SNAPSHOT.jar").setExecutable(false).setCache(false).build()
> >> > > > > > > >
> >> > > > > > > > val iuwtExec = "java -jar
> >> > > > > > > > iuwtExecutor-assembly-0.1-SNAPSHOT.jar
> >> > > > > >
> >> > > > > > -
> >> > > > > >
> >> > > > > > > > Xmx1024M -Xmx128M"
> >> > > > > > > >
> >> > > > > > > > val iuwtCommand =
> >> > > > > > > >
> >> > > > > > > > CommandInfo.newBuilder.setValue(iuwtExec).addAllUris(
> >> > >
> >> > > List(iuwtjURI,
> >> > >
> >> > > > > > > > iuwtURI).asJava).setShell(true).build()
> >> > > > > > > >
> >> > > > > > > > val iuwtImageInfo =
> >> > > > > > > >
> >> > > > > > > > Image.newBuilder().setType(Image.Type.DOCKER).setDocker(
> >> > > > > > > > Image.Docker.newBuilder.setName("ubuntu-
> >> > > > > > > > mesos:0.11-17102016-IUWT").build()).build()
> >> > > > > > > >
> >> > > > > > > > val iuwtContInfo =
> >> > > > > > > >
> >> > > > > > > > ContainerInfo.MesosInfo.newBuilder().setImage(
> >> > >
> >> > > iuwtImageInfo).build()
> >> > >
> >> > > > > > > > val iuwtContainer = ContainerInfo.newBuilder()
> >> > > > > > > >
> >> > > > > > > > .setMesos(iuwtContInfo)
> >> > > > > > > >
> >> > > > > > > > .setType(ContainerInfo.Type.MESOS)
> >> > > > > > > > .build()
> >> > > > > > > >
> >> > > > > > > > val iuwtExecutor = ExecutorInfo.newBuilder()
> >> > > > > > > >
> >> > > > > > > > .setCommand(iuwtCommand)
> >> > > > > > > > .setContainer(iuwtContainer)
> >> > > > > > > > .setExecutorId(ExecutorID.
> >> > >
> >> > > newBuilder().setValue("iuw
> >> > >
> >> > > > > > > > t-
> >> > > > > > > >
> >> > > > > > > > executor"))
> >> > > > > > > >
> >> > > > > > > > .setName("iuwt-executor").build()
> >> > > > > > > >
> >> > > > > > > > The ubuntu-mesos:0.11-17102016-IUWT is a locally hosted
> >>
> >> docker
> >>
> >> > > > > > > > image.
> >> > > > > > > > IUWT is
> >> > > > > > > > the program I'm trying to run, and it runs perfectly fine
> >>
> >> in the
> >>
> >> > > > > > > > aforementioned docker container when running on docker.
The
> >> > > > > >
> >> > > > > > libtiff.so.5
> >> > > > > >
> >> > > > > > > > problem only manifests when I'm using mesos' containerizer.
> >> > > > > > > >
> >> > > > > > > > On Monday, October 17, 2016 6:52:17 AM CEST Avinash
> >>
> >> Sridharan
> >>
> >> > > wrote:
> >> > > > > > > > > You are running a container with its own image right? So
> >>
> >> is
> >>
> >> > > > > > > > /usr/lib/x86_64
> >> > > > > > > >
> >> > > > > > > > > in the container's file system or the host file system?
> >> > > > > > > > >
> >> > > > > > > > > On Mon, Oct 17, 2016 at 6:47 AM, Mark Hammons <
> >> > > > > > > >
> >> > > > > > > > mark.hamm...@inaf.cnrs-gif.fr
> >> > > > > > > >
> >> > > > > > > > > > wrote:
> >> > > > > > > > > >
> >> > > > > > > > > > Yes, it's installed under /usr/lib/x86_64 or whatever
> >>
> >> the
> >>
> >> > > > > > > > > > multilib
> >> > > > > > > >
> >> > > > > > > > path is
> >> > > > > > > >
> >> > > > > > > > > > in
> >> > > > > > > > > > debian. It seems files under this path are not
> >>
> >> accessible.
> >>
> >> > > > > > > > > > I added LD_PRELOAD=/usr/lib/libtiff.so.5 to try to
> >>
> >> force the
> >>
> >> > > > > > symlink
> >> > > > > >
> >> > > > > > > > to
> >> > > > > > > >
> >> > > > > > > > > > load
> >> > > > > > > > > > but it refused to load it. I think the mesos
> >>
> >> containerizer
> >>
> >> > > > > > > > > > is
> >> > > > > > > >
> >> > > > > > > > preventing
> >> > > > > > > >
> >> > > > > > > > > > the
> >> > > > > > > > > > program from accessing anything in a directory under
> >> > >
> >> > > /usr/lib/
> >> > >
> >> > > > > > > > > > for
> >> > > > > > > > > > some
> >> > > > > > > > > > reason, as the same program runs fine in the same
> >>
> >> container
> >>
> >> > > > > > > > > > running
> >> > > > > > > >
> >> > > > > > > > under
> >> > > > > > > >
> >> > > > > > > > > > docker.
> >> > > > > > > > > >
> >> > > > > > > > > > On Monday, October 17, 2016 6:40:49 AM CEST Avinash
> >> > > > > > > > > > Sridharan
> >> > > > > >
> >> > > > > > wrote:
> >> > > > > > > > > > > Is the library part of the image that you are
> >> > > > > > > > > > > running?
> >> > > > > > > > > > > Also
> >> > > > > > > > > > > you
> >> > > > > > > > > > > might
> >> > > > > > > > > >
> >> > > > > > > > > > need
> >> > > > > > > > > >
> >> > > > > > > > > > > to setup LD_LIBRARY_PATH in your env while launching
> >>
> >> the
> >>
> >> > > image
> >> > >
> >> > > > > > > > > > > so
> >> > > > > > > >
> >> > > > > > > > that
> >> > > > > > > >
> >> > > > > > > > > > the
> >> > > > > > > > > >
> >> > > > > > > > > > > container process knows where to look for the shared
> >> > >
> >> > > object.
> >> > >
> >> > > > > > > > > > > On Mon, Oct 17, 2016 at 5:21 AM, Mark Hammons <
> >> > > > > > > > > >
> >> > > > > > > > > > mark.hamm...@inaf.cnrs-gif.fr
> >> > > > > > > > > >
> >> > > > > > > > > > > > wrote:
> >> > > > > > > > > > > >
> >> > > > > > > > > > > > Hi all,
> >> > > > > > > > > > > >
> >> > > > > > > > > > > > I've been working with the mesos containerizer to
> >>
> >> handle
> >>
> >> > > my
> >> > >
> >> > > > > > docker
> >> > > > > >
> >> > > > > > > > > > > > containers
> >> > > > > > > > > > > > recently. I created a docker container that
> >> > > > > > > > > > > > requires
> >> > > > > >
> >> > > > > > libtiff.so.5,
> >> > > > > >
> >> > > > > > > > and
> >> > > > > > > >
> >> > > > > > > > > > the
> >> > > > > > > > > >
> >> > > > > > > > > > > > application within it runs well. But when I try to
> >>
> >> run
> >>
> >> > > > > > > > > > > > it
> >> > > > > >
> >> > > > > > within
> >> > > > > >
> >> > > > > > > > the
> >> > > > > > > >
> >> > > > > > > > > > mesos
> >> > > > > > > > > >
> >> > > > > > > > > > > > containerizer I get an error saying libtiff.so.5
> >>
> >> doesn't
> >>
> >> > > > > > > > > > > > exist.
> >> > > > > > > > > > > >
> >> > > > > > > > > > > > The application is being launched via java's
> >> > > > > > > > > > > > process
> >> > > > > > > > > > > > mechanism
> >> > > > > > > > > > > > from
> >> > > > > > > > > >
> >> > > > > > > > > > inside
> >> > > > > > > > > >
> >> > > > > > > > > > > > a
> >> > > > > > > > > > > > java thread in a custom java executor if that makes
> >>
> >> a
> >>
> >> > > > > > difference.
> >> > > > > >
> >> > > > > > > > > > > > Any idea what could be causing this change in
> >>
> >> behavior?
> >>
> >> > > As
> >> > >
> >> > > > > > > > > > > > you
> >> > > > > >
> >> > > > > > can
> >> > > > > >
> >> > > > > > > > see
> >> > > > > > > >
> >> > > > > > > > > > in
> >> > > > > > > > > >
> >> > > > > > > > > > > > the
> >> > > > > > > > > > > > attached log file, I check /usr/lib, and a symbolic
> >>
> >> link
> >>
> >> > > to
> >> > >
> >> > > > > > > > > > > > /usr/lib/x86..../
> >> > > > > > > > > > > > libtiff.so.5 exists in /usr/lib so the program
> >>
> >> should be
> >>
> >> > > > > > > > > > > > able
> >> > > > > >
> >> > > > > > to
> >> > > > > >
> >> > > > > > > > find
> >> > > > > > > >
> >> > > > > > > > > > and
> >> > > > > > > > > >
> >> > > > > > > > > > > > load
> >> > > > > > > > > > > > that....
> >> > > > > > > > > > > > ----
> >> > > > > > > > > > > > Mark Edgar Hammons II | +33 06 03 69 56 56
> >> > > > > > > > > >
> >> > > > > > > > > > --
> >> > > > > > > > > > ----
> >> > > > > > > > > > Mark Edgar Hammons II | +33 06 03 69 56 56
> >> > > > > > > >
> >> > > > > > > > --
> >> > > > > > > > ----
> >> > > > > > > > Mark Edgar Hammons II | +33 06 03 69 56 56
> >> > > > > >
> >> > > > > > --
> >> > > > > > ----
> >> > > > > > Mark Edgar Hammons II | +33 06 03 69 56 56
> >> > >
> >> > > --
> >> > > ----
> >> > > Mark Edgar Hammons II | +33 06 03 69 56 56
> >>
> >> --
> >> ----
> >> Mark Edgar Hammons II | +33 06 03 69 56 56
> >
> > --
> > Avinash Sridharan, Mesosphere
> > +1 (323) 702 5245
--
----
Mark Edgar Hammons II | +33 06 03 69 56 56
signature.asc
Description: This is a digitally signed message part.
