On Wed, Jan 11, 2017 at 4:25 PM, Brian Stansberry <
brian.stansberry(a)redhat.com> wrote:
> On Jan 11, 2017, at 3:38 PM, Stuart Douglas <stuart.w.douglas(a)gmail.com>
wrote:
>
> Ah, I guess you could just install an earlier DUP that attaches the
DEPLOYMENT_ROOT yourself, then DeploymentRootMountProcessor will not run.
>
> I'm not sure about the deployment scanner though. I think if you have a
.dodeploy file it can attempt to deploy anything, however I can't see any
way to make it deploy arbitrary files without a marker file.
>
Perhaps that is for the best. I was thinking about doing the capability
driven thing you mentioned so people can register other types. But the
primary point of registering other types would be to register some “file is
complete” check handler that could be used to confirm that the file is
complete. Without this deploying without a marker is not safe as the
scanner can read a partially copied file. I’m not sure any “file is
complete” check handler is practical for other file types, and if not just
using the marker is the correct behavior.
Perhaps we could just count on java.nio.channels.FileLock for that, but
that’s documented as only being advisory and working with a locked file has
other limitations.[1]
[1]
https://docs.oracle.com/javase/8/docs/api/java/nio/
channels/FileLock.html#pdep
This is advisory, because it depends on a specific POSIX API on LInux/Unix
systems, that may or may not be there depending on the underlying file
system. We ran into this with HornetQ, since it relies on this for its
journal lock for HA. The GFS 2 file system didn't have this lock
so HA for HornetQ didn't work. We had to implement a native code
workaround to use the BSD based API that was functionally the same as the
POSIX API. Eventually the GFS team did add the POSIX API. It turned out,
all the other file systems that have been tested all worked and had the
underlying POSIX API support (including NFS).
Andy
> Stuart
>
> On Thu, Jan 12, 2017 at 8:09 AM, Bob McWhirter <bmcwhirt(a)redhat.com>
wrote:
> TorqueBox allows deployment of .yaml files. It's pluggable enough.
>
> Bob
>
> On Wed, Jan 11, 2017 at 3:32 PM Stuart Douglas <
stuart.w.douglas(a)gmail.com> wrote:
> The problem with that is that the majority of the time this would be a
mistake on the users part (at the moment it is 100% of the time, as we only
deploy stuff we know how to process).
>
> I think that ideally this should be pluggable. It would be easy enough
to make DeploymentRootMountProcessor handle different file types (for
instance add an AttachmentList of known REAL file types, and then install a
DUP before it to add .ddl to the list).
>
> Using capabilities and requirements we could probably add something
similar to the deployment scanner as well, basically your subsystem would
look for the deployment scanner capability, and if it is installed we could
have some API to add additional file types (although this may be a bit
racey, as the service may not know all file types for the initial scan,
which could give odd results in some circumstances. There is probably a
solution to this but I can't think of one off the top of my head).
>
> Stuart
>
>
>
> On Thu, Jan 12, 2017 at 6:51 AM, Ramesh Reddy <rareddy(a)redhat.com>
wrote:
> It would be good if the logic can be changed such that any thing other
than with extension jar,war,ear and zip to consider them as file based
deployments rather than zip archive based deployments.
>
>
>
>
>
> ----- Original Message -----
>
>
> > I’ll defer to our deployment processing gurus re the mount question.
>
>
> >
>
>
> > As an aside though, for this to work with the deployment-scanner
subsystem
>
>
> > we’ll need to add some logic. Right now it would just ignore the file.
>
>
> >
>
>
> > > On Jan 11, 2017, at 8:34 AM, Ramesh Reddy <rareddy(a)redhat.com>
wrote:
>
>
> > >
>
>
> > > Hi,
>
>
> > >
>
>
> > > For Teiid project (
http://teiid.org), we typically deploy a .xml or
.VDB
>
>
> > > (zip archive) file to define a virtual database artifact. We are
planning
>
>
> > > to deliver a feature where a virtual database is written in DDL, for
this
>
>
> > > we would like to deploy a file artifact like "foo-vdb.ddl".
>
>
> > >
>
>
> > > I have written deployment processors for it, and added
DEPLOYMENT_ROOT
>
>
> > > mounter to recognize the deployment artifact etc, however during the
>
>
> > > deployment scanning, WildFly always looks at anything other than
".xml"
>
>
> > > file as zip archive, or a exploded zip archive, so that it can do VFS
>
>
> > > mount on that file. I would like to add this ".ddl" extension
file
exactly
>
>
> > > similar to ".xml" file. Is there any way to achieve this?
>
>
> > >
>
>
> > > Thank you.
>
>
> > >
>
>
> > > Ramesh..
>
>
> > > _______________________________________________
>
>
> > > wildfly-dev mailing list
>
>
> > > wildfly-dev(a)lists.jboss.org
>
>
> > >
https://lists.jboss.org/mailman/listinfo/wildfly-dev
>
>
> >
>
>
> > --
>
>
> > Brian Stansberry
>
>
> > Manager, Senior Principal Software Engineer
>
>
> > JBoss by Red Hat
>
>
> >
>
>
> >
>
>
> >
>
>
> >
>
>
> > _______________________________________________
>
>
> > wildfly-dev mailing list
>
>
> > wildfly-dev(a)lists.jboss.org
>
>
> >
https://lists.jboss.org/mailman/listinfo/wildfly-dev
>
>
>
>
>
> _______________________________________________
>
>
> wildfly-dev mailing list
>
>
> wildfly-dev(a)lists.jboss.org
>
>
>
https://lists.jboss.org/mailman/listinfo/wildfly-dev
>
>
>
> _______________________________________________
>
> wildfly-dev mailing list
>
> wildfly-dev(a)lists.jboss.org
>
>
https://lists.jboss.org/mailman/listinfo/wildfly-dev
>
> _______________________________________________
> wildfly-dev mailing list
> wildfly-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/wildfly-dev
--
Brian Stansberry
Manager, Senior Principal Software Engineer
JBoss by Red Hat
_______________________________________________
wildfly-dev mailing list
wildfly-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/wildfly-dev
--
Andrig (Andy) T. Miller
Global Platform Director, Middleware
Red Hat, Inc.