Why not just set permissions and ownership in your deployment script?
I use bash scripts for deployment and the enabler is my machine root
user having authorized ssh key access to root on the deployment
servers. Hence with a script I can build, rename, copy to remote, set
owner/permissions on remote. When that's all done I go into Monitor
and change config to new named woa for new instances, launch a new
instance and refuse sessions on old instance.
For example on a project called omega, the dist that is built locally
on my machine is automatically named to sth like
omega_20070412_1415.woa before copying to appserver (that's
project_YYYYMMDD_HHMM.woa). Just change config to select the omega
executable inside that after the deployment is done.
HTH, Kieran
On Apr 12, 2007, at 12:19 PM, Fabian Peters wrote:
> Hi,
>
> I looked into using maven-generated archives for actual deployment.
> I'd like to pull the packaged build from an internal maven
> repository on the build server, unpack it and then gracefully
> restart the app via JavaMonitor.
> However, all files (not directories...) get mode 0644 when
> unpacking the archive. Is this intended? Can it be changed? I could
> of course make the files in question executable after extraction,
> but I'd much prefer not having to this.
>
> How do others proceed with deployment? Do you build locally on the
> app-server machine?
>
> cheers
>
> Fabian
This archive was generated by hypermail 2.0.0 : Thu Apr 12 2007 - 12:50:47 EDT