On 13 April 2011 10:02, Alessandra Forti <[log in to unmask]> wrote:
> Just for the record I found missing files using the replication tool (i.e.
> in the db but not on disk). Is the fsck tool going to do this or is it
> something else?
Yes, it will do that (in fact, the replication tool is based on the
shell of the fsck tool, which is why it still complains about missing
files in the way that the fsck tool will).
Sam
>
> cheers
> alessandra
>
> On 13/04/2011 09:25, Sam Skipsey wrote:
>>
>> On 13 April 2011 08:40, Alessandra Forti<[log in to unmask]> wrote:
>>>
>>> Hi Sam,
>>>>
>>>> Here you go.
>>>>
>>> can you make it more standard (add a help or whatever standard option you
>>> have) and add it to the tools? I'm sure someone else will find it useful.
>>
>> I was going to (at the same time I add the dpm-fsck tools for Ewan...
>> which, yes, Ewan, I've not sent to you yet. By the end of today, I
>> promise - the first way I tried doing the reverse check turned out to
>> be horribly inefficient, so I had to recode that.)
>>
>>>> Invocation is
>>>> python dpm-sql-dpns-by-replication.py #
>>>> (or if you set it executable you can execute it as normal)
>>>> where # is the minimum number of replicas you want to consider
>>>> reporting.
>>>
>>> It'd be good if this was an option.
>>>
>> Certainly - my argument for it being an argument was that the purpose
>> of the tool was to report very replicated files, hence specification
>> of a minimum replica count was mandatory. (After all, most of the
>> files in your DPM will have 1 replica only.)
>> I could make it an option and default it to 2 ?
>>
>>> I might start to apply changes myself if I find the time. Would there be
>>> any
>>> objection?
>>>
>> Of course not.
>>
>>> thanks
>>>
>>> cheers
>>> alessandra
>>>
>>>
>
>
|