Hello Petr,
I have been assuming all along that our discussion concerns
only the RedHat distribution spec files. My last comment was
to tell you that if you do not have similar code as the community
project spec files for creating initial random passwords at install
time, you are perhaps providing an installation that is not secure.
I leave it to you to know if my suggestion is interesting and applicable
to your spec file. My desire is to see all installations of Bacula by
various
Linux distros as secure and stable as possible.
Best regards,
Kern
On 03/04/2013 12:40 PM, Petr Hracek wrote:
> Hello Kern,
>
> I think that those discussion is not relevant to bacula-devel.
> Spec file correction which I have delivered is only distribution relevant.
>
> Thanks for cooperation and you time.
>
> BR
> Petr
> On 03/04/2013 10:55 AM, Kern Sibbald wrote:
>> Hello,
>>
>> Thanks for the update. Nice. This seems to me to be the best way
>> to handle the issue.
>>
>> I haven't looked at your full spec file for quite some time and since
>> it is different from the one in the source code, I will just mention that
>> in the last one I saw, the RedHat package did not have the necessary
>> code to automatically generate unique random passwords at initial
>> installation time. You may already have this, but if not, it is relatively
>> easy to do and improves security for users considerable. The project
>> spec files (now probably quite out of date) with the necessary code
>> can be found in <bacula>/platforms/rpms/redhat/bacula.spec.in.
>> In particular search for XXX_REPLACE and you will see how we
>> do it.
>>
>> One thing I didn't like about spec files is that we had to duplicate similar
>> code in a number of places, and I saw no way of creating a sort of
>> macro or function inside the spec file (I guess it can be done by some
>> sort of included files). If you have a better way of resolving that issue
>> please let me know -- I survive, but I am far from being good at writing
>> rpm spec files.
>>
>> Best regards,
>> Kern
>>
>> On 03/04/2013 10:33 AM, Petr Hracek wrote:
>>> Hello,
>>>
>>> Yes I think that separate package would be good in spec file
>>> I have made modifications against upstream:
>>> diff --git a/bacula.spec b/bacula.spec
>>> index cc5c31c..3bd79ad 100644
>>> --- a/bacula.spec
>>> +++ b/bacula.spec
>>> @@ -212,9 +212,50 @@ verification of computer data across a network of
>>> different computers. It is
>>> based on a client/server architecture.
>>>
>>> This package contains the storage daemon, the daemon responsible for
>>> writing
>>> -the data received from the clients onto tape drives or other mass storage
>>> -devices.
>>> +the data received from the clients onto other mass storage devices.
>>>
>>> +%package storage-tape
>>> +Summary: Bacula storage daemon files
>>> +Group: System Environment/Daemons
>>> +Requires: bacula-common%{?_isa} = %{version}-%{release}
>>> +# Storage backends merged into core.
>>> +Provides: bacula-storage-common = %{version}-%{release}
>>> +Obsoletes: bacula-storage-common < 5.2.2-2
>>> +Provides: bacula-storage-mysql = %{version}-%{release}
>>> +Obsoletes: bacula-storage-mysql < 5.2.0
>>> +Provides: bacula-storage-sqlite = %{version}-%{release}
>>> +Obsoletes: bacula-storage-sqlite < 5.2.0
>>> +Provides: bacula-storage-postgresql = %{version}-%{release}
>>> +Obsoletes: bacula-storage-postgresql < 5.2.0
>>> +Requires: mt-st
>>> +
>>> +%if 0%{?fedora} == 17
>>> +Requires(post): systemd-units
>>> +Requires(preun): systemd-units
>>> +Requires(postun): systemd-units
>>> +%endif
>>> +
>>> +%if 0%{?fedora} >= 18 || 0%{?rhel} >= 7
>>> +Requires(post): systemd-sysv
>>> +Requires(post): systemd
>>> +Requires(preun): systemd
>>> +Requires(postun): systemd
>>> +%endif
>>> +
>>> +%if 0%{?rhel} == 5 || 0%{?rhel} == 6
>>> +Requires(post): /sbin/chkconfig
>>> +Requires(preun): /sbin/chkconfig
>>> +Requires(preun): /sbin/service
>>> +Requires(postun): /sbin/service
>>> +%endif
>>> +
>>> +%description storage-tape
>>> +Bacula is a set of programs that allow you to manage the backup,
>>> recovery, and
>>> +verification of computer data across a network of different computers.
>>> It is
>>> +based on a client/server architecture.
>>> +
>>> +This package contains the storage daemon, the daemon responsible for
>>> writing
>>> +the data received from the clients onto tape drives devices.
>>>
>>> %package client
>>> Summary: Bacula backup client
>>> @@ -844,6 +885,11 @@ fi
>>> %{_mandir}/man8/bscan.8.*
>>> %{_mandir}/man8/btape.8.*
>>>
>>> +%files storage-tape
>>> +%defattr(-,root,root,-)
>>> +%{_libexecdir}/%{name}/mtx-changer
>>> +%{_libexecdir}/%{name}/mtx-changer.conf
>>> +
>>> %files client
>>> %defattr(-,root,root,-)
>>> %{_sbindir}/bacula-fd
>>>
>>> [stone@kiasportyw 5.fc19(master)]$ rpm -qpl
>>> bacula-storage-tape-5.2.13-5.fc19.i686.rpm
>>> /usr/libexec/bacula/mtx-changer
>>> /usr/libexec/bacula/mtx-changer.conf
>>> [stone@kiasportyw 5.fc19(master)]$ rpm -qpl --requires
>>> rpm: no arguments given for query
>>> [stone@kiasportyw 5.fc19(master)]$ rpm -qpl --requires
>>> bacula-storage-tape-5.2.13-5.fc19.i686.rpm
>>> /bin/sh
>>> bacula-common(x86-32) = 5.2.13-5.fc19
>>> mt-st
>>> rpmlib(CompressedFileNames) <= 3.0.4-1
>>> rpmlib(FileDigests) <= 4.6.0-1
>>> rpmlib(PayloadFilesHavePrefix) <= 4.0-1
>>> systemd
>>> systemd
>>> systemd
>>> systemd-sysv
>>> rpmlib(PayloadIsXz) <= 5.2-1
>>> /usr/libexec/bacula/mtx-changer
>>> /usr/libexec/bacula/mtx-changer.conf
>>> [stone@kiasportyw 5.fc19(master)]$
>>>
>>> Would it be OK?
>>>
>>> On 03/01/2013 11:37 PM, Kern Sibbald wrote:
>>>> Hello,
>>>>
>>>> Yes, I have been aware of the dependency problem for
>>>> some time. The problem is that if the user wants only a disk
>>>> based backup, which is becoming more and more prevalent,
>>>> such a dependency forces the user to load a package that
>>>> may never be used.
>>>>
>>>> Possibly the right way to do it is to have a separate Bacula
>>>> package that adds tape backup.
>>>>
>>>> Nothing really pleases me, so in such situations, I usually
>>>> do nothing until the right solution arises.
>>>>
>>>> What do you think?
>>>>
>>>> Best regards,
>>>> Kern
>>>>
>>>> On 03/01/2013 10:56 AM, Petr Hracek wrote:
>>>>> Hi,
>>>>>
>>>>> on the base of that bug:
>>>>> https://bugzilla.redhat.com/show_bug.cgi?id=911573
>>>>>
>>>>> Means that when mt command is not existing on the system them mtx-change
>>>>> does not work.
>>>>> /usr/libexec/bacula/mtx-changer
>>>>> or causes problem.
>>>>>
>>>>> I have prepared patch which should solve that problem.
>>>>>
>>>>> diff --git a/bacula.spec b/bacula.spec
>>>>> index 8d518a0..a13d35e 100644
>>>>> --- a/bacula.spec
>>>>> +++ b/bacula.spec
>>>>> @@ -185,6 +185,7 @@ Provides: bacula-storage-sqlite =
>>>>> %{version}-%{release}
>>>>> Obsoletes: bacula-storage-sqlite < 5.2.0
>>>>> Provides: bacula-storage-postgresql =
>>>>> %{version}-%{release}
>>>>> Obsoletes: bacula-storage-postgresql < 5.2.0
>>>>> +Requires: mt-st
>>>>>
>>>>> %if 0%{?fedora} == 17
>>>>> Requires(post): systemd-units
>>>>>
>>>>> What do you think about that?
>>>>>
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb
_______________________________________________
Bacula-devel mailing list
Bacula-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-devel
