Hi All,
This makes more sense then. To update dpm I just did:
yum update dpm
but it seems I needed to do:
yum update DPM-server-mysql
as well because otherwise the script wasn't happy. Maybe I should have
done a full yum update instead...
Thanks,
Mark
On 17/05/2012 12:04, Sam Skipsey wrote:
> ...no, it's sillier than that.
> I do have the dpm-server-mysql metapackage installed... but it's
> capitalised like that (not like DPM-server-mysql for 1.8.2 and
> earlier). Obviously, this doesn't match the search done by the script,
> so the version remains unset.
>
> I shall file a bug.
>
> Sam
>
> On 17 May 2012 12:02, Sam Skipsey<[log in to unmask]> wrote:
>> ah, yes, that'll be the problem. I don't have the metapackage
>> installed, due to the various package jiggering about that happened
>> debugging the 1.8.3 crash bug.
>>
>> :D
>>
>> Sam
>>
>> On 17 May 2012 11:59, Ewan MacMahon<[log in to unmask]> wrote:
>>>> -----Original Message-----
>>>> From: GRIDPP2: Deployment and support of SRM and local storage management
>>>> [mailto:[log in to unmask]] On Behalf Of Sam Skipsey
>>>> Sent: 17 May 2012 11:51
>>>>
>>>> ...yes. In the python file "dpm-listspaces", the code explicitly sets the
>>>> value of "version" for the DPM to "unset".
>>>>
>>> This is (I think) the relevant fragment from my EMI 1.8.2, so it
>>> unsets it, then sets it based on an RPM query. Which, apparently,
>>> works 'normally'.
>>>
>>> self.version = 'unset'
>>> if self.name == socket.getfqdn():
>>> self.core_services.append(StorageService('dpm'))
>>> self.core_services.append(StorageService('dpns'))
>>>
>>> ts = rpm.TransactionSet()
>>> mi = ts.dbMatch()
>>> mi.pattern('name',rpm.RPMMIRE_GLOB,'DPM-server-*')
>>> # Normally only one RPM will match...
>>> for package in mi:
>>> self.version = package['version']
>>>
>>> Ewan
|