[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: conflict between seedit <-> selinux-policy and qstat <-> torque-client



On Wed, Nov 4, 2009 at 4:43 PM, Rudolf Kastl <che666 gmail com> wrote:
> 2009/11/4 Steve Traylen <steve traylen cern ch>:
>> On Wed, Nov 4, 2009 at 4:33 PM, Rudolf Kastl <che666 gmail com> wrote:
>>> 2009/11/4 Steve Traylen <steve traylen cern ch>:
>>>> On Wed, Nov 4, 2009 at 4:11 PM, Rudolf Kastl <che666 gmail com> wrote:
>>>>> 2009/11/4 Jason L Tibbitts III <tibbs math uh edu>:
>>>>>>>>>>> "ST" == Steve Traylen <steve traylen cern ch> writes:
>>>>>>
>>>>>> ST> Would be happy for an alternatives solution. I have yet another
>>>>>> ST> /usr/bin/qstat for a POSIX interface to batch on the way at some
>>>>>> ST> point.
>>>>>>
>>>>>> Turns out that the other queuing systems (torque and gridengine) have
>>>>>> already renamed their qstat binaries (to qstat-torque and qstat-ge).  I
>>>>>> would expect that other queuing packages should do the same.
>>>>>
>>>> Yes a qstat-slurm with qstat as alternative across them.
>>>> Good news.
>>>
>>> but then the alternatives qstat  conflicts with /usr/bin/qstat from
>>> the qstat rpm package, doesent it?
>>
>> The torque  spec is creating correctly /usr/bin/qstat as a symlink
>> via alternatives mechanism (reading the .spec only, have not checked).
>>
>> The qstat pkg should do the same. Currently while the qstat
>> pkg is creating a file at /usr/bin/qstat then it is conflicting in
>> the RPM sense. Once qstat pkg uses alternatives as well
>> it will no longer conflict.
>>
>> Two packages that contain alternatives for a single file
>> don't conflct in the RPM sense. You can install both pkgs
>> and then select one to be the real /usr/bin/qstat via
>> the alternatives mechanism.
>> Hope that makes sense.
>
> it does with one exception... the qstat rpm is basically "quake stat".
> so it does something completly different than the qstat of torque or
> gridengine and hmm the real resolution would maybe be to rename the
> binary of the qstat package then.
Yes you are right, the qstat qstat package should rename since that location
is reserved by POSIX and anyway torque got there first.


>
> kind regards,
> Rudolf Kastl
>
> p.s. thanks everyone for the replies and the effort done already.
>
>>
>> Steve
>>
>>
>>>
>>> kind regards,
>>> Rudolf Kastl
>>>
>>>>
>>>>> that means that the conflict tags in the qemu and the torque-clients
>>>>> package are invalid.
>>>>>
>>>>> thanks for checking jason!
>>>>>
>>>>> kind regards,
>>>>> Rudolf Kastl
>>>>>>
>>>>>>  - J<
>>>>>>
>>>>>> --
>>>>>> fedora-devel-list mailing list
>>>>>> fedora-devel-list redhat com
>>>>>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>>>>>>
>>>>>
>>>>> --
>>>>> fedora-devel-list mailing list
>>>>> fedora-devel-list redhat com
>>>>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Steve Traylen
>>>>
>>>> --
>>>> fedora-devel-list mailing list
>>>> fedora-devel-list redhat com
>>>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>>>>
>>>
>>> --
>>> fedora-devel-list mailing list
>>> fedora-devel-list redhat com
>>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>>>
>>
>>
>>
>> --
>> Steve Traylen
>>
>> --
>> fedora-devel-list mailing list
>> fedora-devel-list redhat com
>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>>
>
> --
> fedora-devel-list mailing list
> fedora-devel-list redhat com
> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>



-- 
Steve Traylen


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]