what is the status of GRIO support in the Linux port of XFS?
Also, if the answer is 'non existent', what is the recommended
alternative? I've got an application that needs to stream a
huge amount of data to the harddrive without dropping any and
without blocking the sender. We will be pushing the limits
of our high-end raid striped disks. This seems the exactly
the type of thing GRIO was made for, but last I heard it was
missing from Linux XFS with no plans to add it. Any change
in that? I know I can get almost there with I/O priorities
and the RT features in 2.6... but its not quite the same
thing.
Apologies if this has been beat to death here or elsewhere...
I've googled the heck out of this and rummaged around in the
list archives (as much as this fscking corporate firewall will
let me) with little result. I'll gladly RTFM if someone can
point me at the right one. :-/
Feel free to CC me on replies, as I read the LKML in digest
format.
Thanks,
Thad Phetteplace
>
>what is the status of GRIO support in the Linux port of XFS?
Called realtime volume.
(http://en.wikipedia.org/wiki/XFS section 2.11)
>Also, if the answer is 'non existent', what is the recommended
>alternative? I've got an application that needs to stream a
>huge amount of data to the harddrive without dropping any and
>without blocking the sender. We will be pushing the limits
>of our high-end raid striped disks. This seems the exactly
>the type of thing GRIO was made for, but last I heard it was
>missing from Linux XFS with no plans to add it. Any change
>in that? I know I can get almost there with I/O priorities
>and the RT features in 2.6... but its not quite the same
>thing.
>
>Apologies if this has been beat to death here or elsewhere...
>I've googled the heck out of this and rummaged around in the
>list archives (as much as this fscking corporate firewall will
>let me) with little result. I'll gladly RTFM if someone can
>point me at the right one. :-/
>
>Feel free to CC me on replies, as I read the LKML in digest
>format.
>
>Thanks,
>
>Thad Phetteplace
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to [email protected]
>More majordomo info at http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at http://www.tux.org/lkml/
>
-`J'
--
Jan Engelhardt wrote:
>> what is the status of GRIO support in the Linux port of XFS?
>
> Called realtime volume.
> (http://en.wikipedia.org/wiki/XFS section 2.11)
Well, not really. realtime was a part of the old griov1 setup on Irix,
but realtime != GRIO.
-Eric
On Wed, 2006-10-11 at 16:16 -0500, Eric Sandeen wrote:
> Jan Engelhardt wrote:
> >> what is the status of GRIO support in the Linux port of XFS?
> >
> > Called realtime volume.
> > (http://en.wikipedia.org/wiki/XFS section 2.11)
>
> Well, not really. realtime was a part of the old griov1 setup on Irix,
> but realtime != GRIO.
the realtime allocator really should be renamed bitmap.
griov1 relied on an fs allocator that would return in bounded time
so it could satisfy irix realtime requirements thus it was
called "realtime".
> -Eric
>
--
Russell Cattelan <[email protected]>
Jan Engelhardt wrote:
>> what is the status of GRIO support in the Linux port of XFS?
>>
griov2 supports XFS Linux/Irix on local and cluster volumes but it is
not an open source
project. Please check this link for more information:
http://techpubs.sgi.com/library/tpl/cgi-bin/getdoc.cgi/0650/bks/SGI_Admin/books/GRIO2_AG/sgi_html/ch01.html
>
> Called realtime volume.
> (http://en.wikipedia.org/wiki/XFS section 2.11)
>
>
>> Also, if the answer is 'non existent', what is the recommended
>> alternative? I've got an application that needs to stream a
>> huge amount of data to the harddrive without dropping any and
>> without blocking the sender. We will be pushing the limits
>> of our high-end raid striped disks. This seems the exactly
>> the type of thing GRIO was made for, but last I heard it was
>> missing from Linux XFS with no plans to add it. Any change
>> in that? I know I can get almost there with I/O priorities
>> and the RT features in 2.6... but its not quite the same
>> thing.
>>
>> Apologies if this has been beat to death here or elsewhere...
>> I've googled the heck out of this and rummaged around in the
>> list archives (as much as this fscking corporate firewall will
>> let me) with little result. I'll gladly RTFM if someone can
>> point me at the right one. :-/
>>
>> Feel free to CC me on replies, as I read the LKML in digest
>> format.
>>
>> Thanks,
>>
>> Thad Phetteplace
>>
>> -
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to [email protected]
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at http://www.tux.org/lkml/
>>
>>
>
> -`J'
>