[Kolab-devel] bash in gcc build can not find stage2/xgcc

[Bernhard Reiter]

On Wednesday 05 January 2005 22:36, Richard Bos wrote:
> Op woensdag 5 januari 2005 15:17, schreef Martin Konold:
> > Please contact the OpenPKG developers and report back to us.
>
> Well I just repeated the 'obmtool kolab' command, to get more info and
> guess what all went fine....  Nothing changed as far as I know.

Strange,
it still makes sense to contact the openpkg people.

> In the meantime I have successfully build all the packages on suse-9.2
> But I had to repeat the obmtool many times, because needed packages where
> not build in time and because of that the other packages were not build....
> E.g. grep needs pcre, but the latter is build much later than grep.  Once
> pcre has been build, it is possible to build grep.  Once that one is build
> some dependend perl modules can be build (if I remember correctly).
>
> I have now 76 packages installed:
> # /kolab/bin/openpkg rpm -qa | wc -l
> 76
>
> Is that correct?  It would be nice if this mentioned in the wiki, perhaps
> with the full list of packages.

The full list of packages is inside of the obmtool.conf.

> Further more the rquired filespace is about 775 MB:
> # du -sh /srv/kolab-server-2.0-beta-1/
> 775M    /srv/kolab-server-2.0-beta-1/
>
> Same as above, it would be nice if this is mentioned on the wiki (I run out
> of filespace during the build, resulting in yet another obmtool call ;)

Can you add it?
This is a wiki for a reason after all ;-)


> Waiting for OpenLDAP to start
> could not connect ldap server ldap://127.0.0.1:389/
> at /kolab/etc/kolab/kolab_bootstrap line 408.
> Can't call method "bind" without a package or object reference
> at /kolab/etc/kolab/kolab_bootstrap line 410.
>
> ------------------
> What can be the reason that the bootstrapper can't connect to the ldap
> server (To be honest I find it a cryptic error message ;)

I could be a firewall setting.
Try starting only openldap and to connect with telnet to
see if this works.  telnet localhost 389

> I assume the error below is, because the bootstrapper did not finish.  I
> get the error during system boot as well as /etc/init.rc/kolab is called
> for the rcN directories.
>
> chaos:~ # /kolab/bin/openpkg rc all start
> OpenPKG: start: openpkg, openldap, sasl, spamassassin, amavisd:FAILED
> openpkg:rc:WARNING: /kolab:amavisd:%start: failed with return code 9
> openpkg:rc:NOTICE: output from stdout/stderr is following:
> +----------------------------------------------------------------------
>
> | No $LOGFILE is specified (and not logging via syslog) at
> | /kolab/sbin/amavisd
>
> line 1050.
> +----------------------------------------------------------------------
> OpenPKG: start: apache, clamav, dcron, imapd, postfix, proftpd, kolabd.

Looks like this is only a warning.
Steffen: Any idea where it comes from?

> Is it possible to reuse the binary rpms that I have now?  Can I use those
> to install on another system?  If so how should that be done?

Yes, yes.
Copy all needed (see obmtool.conf) in a directory and add obmtool
and obmtool conf.
Then you can just start it.
This is just like the binary distribution based on Debian that we do.
Probably documented somewhere within openpkg.

> The wiki is does not fill this gap yet:
> http://kodoc.subvs.co.uk/index.php/Kolab2_Installation_-_Binary

Feel free to improve it. :)