cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rafael Fonseca <>
Subject Re: Cloustack packaging
Date Sun, 26 Apr 2015 14:34:42 GMT
Hi Remi,

I'm building the RPMs from the latest master branch, i just referred to
that old branch because it was the only one i found where someone was
working on this.

Some of the issues i've encountered were:

Thsese files did not get created by the install, so
the cloudstack-setup-management script failed.

Also, when this script failed it refered to a wrong log file...
Configure CloudStack Management Server ...[Failed]
Failed to configure CloudStack Management Server, please see the
/var/log/cloudstack/setupManagement.log for detail
wrong file.. correct is /var/log/cloudstack/management/setupManagement.log

Besides that, the systemd init script tries to use /usr/sbin/tomcat-sysd to
start the service.. this file doesn't exist, only /usr/sbin/tomcat (perhaps
it got changed in a more recent version of the tomcat rpm that is in the
CentOS 7 repos...
Anyway.. the current tomcat systemd script is
calling /usr/libexec/tomcat/server and not any of those.
If i create a symlink from /usr/sbin/tomcat-sysd to /usr/sbin/tomcat it's
still missing the $NAME to be able to start the service.. though i can
start it manually if I do "export NAME=cloudstack-management"

Another thing i spotted in the logs is this:
DEBUG:root:execute:chkconfig --del tomcat6
DEBUG:root:Failed to execute:error reading information on service tomcat6:
No such file or directory

Which causes the cloudstack-setup-management to fail on the first run,
after i put in the missing files.. on the second run it's fine, but the
tomcat service is still registered in systemd ;)

Some more issues are occurring if you uninstall some packages (nothing too
nasty, but still worth looking into) and specifically i had problems with
reinstalling the cloudstack-management package... i didn't look much into
that yet, but it may be related to the server not getting stopped
automatically when i removed the packages... i needed do "rm -rf
/usr/share/cloudstack-management/*" before i could reinstall.

I ran into a couple more issues, but these are the ones i remember off the
top of my head or that i made some notes about.
I'll patch it myself and submit the code for review as soon as i have some
more time, perhaps still later today :)


On Sun, Apr 26, 2015 at 4:11 PM, Remi Bergsma <> wrote:

> Hi Rafael,
> Welcome to CloudStack and thanks for your mail.
> What RPM packages did you use for your testing?
> From my personal experience I know that the CentOS7 (el7) packages for
> CloudStack 4.4 work just fine as a KVM hypervisor (on CentOS 7.1). In 4.5
> work is done to also make it work as a Management server. So, what exact
> problems do you run into when installing the RPM’s on CentOS 7.1?
> As far as I know, the centos7-rpm branch is old and no longer maintained.
> I’d recommend using the 4.5 and/or master branches instead, as these are
> most up to date and actively maintained.
> Feel free to send a PR to fix what you think is broken.
> Regards,
> Remi
> > On 26 Apr 2015, at 15:30 , Rafael Fonseca <> wrote:
> >
> > Hi all,
> >
> > I'm new to the cloudstack community, but i'm a seasoned developer and
> > systems engineer and i'm sure i'll be able to help out a lot :)
> >
> > I've just started on getting cloudstack running on Centos 7.1 and noticed
> > many problems, which make it NOT work out of the box.
> >
> > I'll be doing a series of fixes and submitting the code for review, since
> > i'm not a commiter yet, if no one else is working on this.
> >
> > I've seen that there's a branch called centos7-rpm which is already
> merged
> > with master, but there hasn't been any activity on that on the last six
> > months, so i guess it's currently stale.
> >
> > Perhaps some of these things that i spotted were working on 7.0, but i
> > haven't tested it out on that version yet.. ideally we can make specs
> that
> > work on ANY 7.x version and not have a spec file for every minor release.
> >
> > One of the things i immediately noticed (that is present in all
> > centos/fedora specs, not just centos7) is the following:
> >
> > %define __os_install_post %{nil}
> > %global debug_package %{nil}
> >
> > # DISABLE the post-percentinstall java repacking and line number
> stripping
> > # we need to find a way to just disable the java repacking and line
> number
> > stripping, but not the autodeps
> >
> >
> > It's easy enough to get the __os_install_post to only do that, but it
> > shouldn't be disabling autodeps.. rather the rpmbuild's dependency
> routines
> > can't find any required external libs in the java or python files :)
> >
> > Replace:
> >
> > %define __os_install_post %{nil}
> >
> > with:
> >
> > %define __jar_repack %{nil}
> > %define __debug_package %{nil}
> >
> > That will do it, as checked in "rpm --showrc", although there won't be
> much
> > to be gained with it.
> >
> > -14: __os_install_post
> >    /usr/lib/rpm/redhat/brp-compress
> >    %{!?__debug_package:
> >    /usr/lib/rpm/redhat/brp-strip %{__strip}
> >    /usr/lib/rpm/redhat/brp-strip-comment-note %{__strip} %{__objdump}
> >    }
> >    /usr/lib/rpm/redhat/brp-strip-static-archive %{__strip}
> >    /usr/lib/rpm/brp-python-bytecompile %{__python}
> > %{?_python_bytecompile_errors_terminate_build}
> >    /usr/lib/rpm/redhat/brp-python-hardlink
> >    %{!?__jar_repack:/usr/lib/rpm/redhat/brp-java-repack-jars}
> > %{nil}

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message