Mailing List Archive

"Grüninger, Andreas (LGL Extern)" <Andreas.Grueninger@lgl.bwl.de>
writes:

> I managed to get Hawk running on Solaris.
> I use as prefix /opt and not /usr.
>
> The last problem seems to be the runtime environment of the ruby application.
>
> For testing purposes I started lighthhtpd in a shell in the foreground.
> In this shell crm works without any problem.
> If in hawk a function is used which uses crm the error message occurs when PYTHONPATH is not defined.
> Crmsh is located in /opt/lib/python2.6/site-packages/.
> When I define links into /usr/lib/python2.6/site-packages hawk can us crm.
>
> A similar problem exists with hb_report.
> I checked hb_report in a shell and it works with some minor problems.
>
> Used in hawk I get error messages that egrep can’t be used.
> Hb_report started over hawk uses /usr/bin/grep and that is the wrong version.
> It should use /usr/gnu/bin which is defined in the PATH before /usr/bin.
> The error may also stem from a process using the output of hb_report.
>
> Hawk or ruby seems to ignore the environment of the starting process.
>
> Does anyone have an idea how to solve the problem?

This may be due to which user hawk is being run as. I think by default
hawk runs as the hacluster user, so the PATH would need to be correct
for that user.

>
> Andreas
> _______________________________________________
> Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

--
// Kristoffer Grönlund
// kgronlund@suse.com

_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org

Found it in tools/hawk_invoke.c:

setenv("PATH", SBINDIR":"BINDIR":/bin", 1);

If you use Solaris/Illumos the gnu utilities are in /usr/gnu/bin. And if you use OpenCSW there exists the path /opt/opencsw/bin.
And with pkgsrc the path must contain “/opt/local/abin:/opt/local/bin”. Pkgsrc can be used with Solais/Illumos/MacOS/BSD.

This is valid for all other environment variables which must be available e.g. PYTHONPATH

…
#if defined(__sun)
setenv("PATH", SBINDIR":"BINDIR":"USRDIRS, 1);
setenv("PYTHONPATH",PYTHONPATH, 1);
#else
…
Where USRDIRS is handed over via Makefile.



Von: Grüninger, Andreas (LGL Extern)
Gesendet: Sonntag, 16. November 2014 21:27
An: 'The Pacemaker cluster resource manager'
Betreff: Hawk and runtime environment

I managed to get Hawk running on Solaris.
I use as prefix /opt and not /usr.

The last problem seems to be the runtime environment of the ruby application.

For testing purposes I started lighthhtpd in a shell in the foreground.
In this shell crm works without any problem.
If in hawk a function is used which uses crm the error message occurs when PYTHONPATH is not defined.
Crmsh is located in /opt/lib/python2.6/site-packages/.
When I define links into /usr/lib/python2.6/site-packages hawk can us crm.

A similar problem exists with hb_report.
I checked hb_report in a shell and it works with some minor problems.

Used in hawk I get error messages that egrep can’t be used.
Hb_report started over hawk uses /usr/bin/grep and that is the wrong version.
It should use /usr/gnu/bin which is defined in the PATH before /usr/bin.
The error may also stem from a process using the output of hb_report.

Hawk or ruby seems to ignore the environment of the starting process.

Does anyone have an idea how to solve the problem?

Andreas