* File Server extention for mirrored devices
@ 1997-07-12 2:00 beto
0 siblings, 0 replies; 4+ messages in thread
From: beto @ 1997-07-12 2:00 UTC (permalink / raw)
In <199707111304.IAA07712@ns.dbSystems.com>
"G. David Butler" <gdb@dbSystems.com> wrote:
> I would like to propose the following extension to fsconfig(8):
>
> { device... }
> A pseudo-device formed by mirroring corresponding blocks of the
> devices in the list. The size of the result is the number of
> blocks on the smallest device.
>
Well if we are proposing new devices I'll like to have
stripping on the fs, or even raid.
fs has interleave and concat disks but the
max. performance is limited by one disk. It does not
break a block in pieces and does the IO in parallel.
I don't think it's applicable to the juke box,
a lot of noise in the machine room :-)
^ permalink raw reply [flat|nested] 4+ messages in thread
* File Server extention for mirrored devices
@ 1997-07-13 1:02 Frank
0 siblings, 0 replies; 4+ messages in thread
From: Frank @ 1997-07-13 1:02 UTC (permalink / raw)
------ =_NextPart_000_01BC8F20.15C9BF80
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
David,
I like your idea, and it fits in well with the current syntax.
You may wish to consider closest head not round robin.
I have measured closest head at 50% greater Kps read
when compared to round robin.
Frank
----------
From: G. David Butler[SMTP:gdb@dbSystems.com]
Sent: Saturday, July 12, 1997 3:59 PM
To: 9fans@cse.psu.edu
Subject: Re: File Server extention for mirrored devices
>From: beto@ncube.com
>
>In <199707111304.IAA07712@ns.dbSystems.com>
> "G. David Butler" <gdb@dbSystems.com> wrote:
>
>> I would like to propose the following extension to fsconfig(8):
>>
>> { device... }
>> A pseudo-device formed by mirroring corresponding blocks of the
>> devices in the list. The size of the result is the number of
>> blocks on the smallest device.
>>
This is RAID 1, no?
>Well if we are proposing new devices I'll like to have
>stripping on the fs, or even raid.
>
>fs has interleave and concat disks but the
>max. performance is limited by one disk. It does not
Huh?
>break a block in pieces and does the IO in parallel.
No, but it interleaves the blocks across drives thereby
doing the IO in "parallel".
>I don't think it's applicable to the juke box,
>a lot of noise in the machine room :-)
Why not. You have two (or more) jukes then you can
inteleave between either platters or entire boxes.
The configuration langauge Ken Thompson has created is an
elegant way of specifing file system layouts. Since it
has recursive definitions you can have interleave concat
drives or vice versa.
So if you have scsi disks targets 0 and 1 and juke boxes on targets 2 and 3:
(the amount of scsi disk is dependent on how you plan to use the system.)
I wouldn't recommend this!!
filesys main c[w0.<0-1>][(r0.2.<0-99>r0.3.<0-99>)]
This would be better.
filesys main [cw0.0r0.2.<0-99>cw0.1r0.3.<0-99>]
More conventional.
filesys main c[w0.<0-1>](r0.2.<0-99>r0.3.<0-99>)
If you have it, flaunt it!
filesys main c[w0.<0-1>][r0.2.<0-99>r0.3.<0-99>]
With my proposal you can also have mirrored stripes, or concat mirrors.
filesys main {[w<0-1>][w<2-3>]} mirror of stripe of drives 0,1 and 2,3
filesys main [{w<0-1>}{w<2-3>}] stripe of mirror of drives 0,1 and 2,3
You get the idea.
As far as parity strips (like RAID 5 or 7), I would recommend buying an
external RAID and put it on the filesystem. I don't like them much
since they can perform very poorly in failure modes. Mirrors are more
expensive but do not destroy performance. Also mirrors, as defined
above, can consist of 3 or more members and thus not fail on a two drive
failure. RAID 5 or 7 can't do that.
Also, I am thinking of implementing the mirror so the reads are
round-robined (is that a word?) across the members for increased
performance.
David Butler
gdb@dbSystems.com
------ =_NextPart_000_01BC8F20.15C9BF80
Content-Type: application/ms-tnef
Content-Transfer-Encoding: base64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------ =_NextPart_000_01BC8F20.15C9BF80--
^ permalink raw reply [flat|nested] 4+ messages in thread
* File Server extention for mirrored devices
@ 1997-07-12 14:59 G.David
0 siblings, 0 replies; 4+ messages in thread
From: G.David @ 1997-07-12 14:59 UTC (permalink / raw)
>From: beto@ncube.com
>
>In <199707111304.IAA07712@ns.dbSystems.com>
> "G. David Butler" <gdb@dbSystems.com> wrote:
>
>> I would like to propose the following extension to fsconfig(8):
>>
>> { device... }
>> A pseudo-device formed by mirroring corresponding blocks of the
>> devices in the list. The size of the result is the number of
>> blocks on the smallest device.
>>
This is RAID 1, no?
>Well if we are proposing new devices I'll like to have
>stripping on the fs, or even raid.
>
>fs has interleave and concat disks but the
>max. performance is limited by one disk. It does not
Huh?
>break a block in pieces and does the IO in parallel.
No, but it interleaves the blocks across drives thereby
doing the IO in "parallel".
>I don't think it's applicable to the juke box,
>a lot of noise in the machine room :-)
Why not. You have two (or more) jukes then you can
inteleave between either platters or entire boxes.
The configuration langauge Ken Thompson has created is an
elegant way of specifing file system layouts. Since it
has recursive definitions you can have interleave concat
drives or vice versa.
So if you have scsi disks targets 0 and 1 and juke boxes on targets 2 and 3:
(the amount of scsi disk is dependent on how you plan to use the system.)
I wouldn't recommend this!!
filesys main c[w0.<0-1>][(r0.2.<0-99>r0.3.<0-99>)]
This would be better.
filesys main [cw0.0r0.2.<0-99>cw0.1r0.3.<0-99>]
More conventional.
filesys main c[w0.<0-1>](r0.2.<0-99>r0.3.<0-99>)
If you have it, flaunt it!
filesys main c[w0.<0-1>][r0.2.<0-99>r0.3.<0-99>]
With my proposal you can also have mirrored stripes, or concat mirrors.
filesys main {[w<0-1>][w<2-3>]} mirror of stripe of drives 0,1 and 2,3
filesys main [{w<0-1>}{w<2-3>}] stripe of mirror of drives 0,1 and 2,3
You get the idea.
As far as parity strips (like RAID 5 or 7), I would recommend buying an
external RAID and put it on the filesystem. I don't like them much
since they can perform very poorly in failure modes. Mirrors are more
expensive but do not destroy performance. Also mirrors, as defined
above, can consist of 3 or more members and thus not fail on a two drive
failure. RAID 5 or 7 can't do that.
Also, I am thinking of implementing the mirror so the reads are
round-robined (is that a word?) across the members for increased
performance.
David Butler
gdb@dbSystems.com
^ permalink raw reply [flat|nested] 4+ messages in thread
* File Server extention for mirrored devices
@ 1997-07-11 13:04 G.David
0 siblings, 0 replies; 4+ messages in thread
From: G.David @ 1997-07-11 13:04 UTC (permalink / raw)
I would like to propose the following extension to fsconfig(8):
{ device... }
A pseudo-device formed by mirroring corresponding blocks of the
devices in the list. The size of the result is the number of
blocks on the smallest device.
David Butler
gdb@dbSystems.com
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~1997-07-13 1:02 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1997-07-12 2:00 File Server extention for mirrored devices beto
-- strict thread matches above, loose matches on Subject: below --
1997-07-13 1:02 Frank
1997-07-12 14:59 G.David
1997-07-11 13:04 G.David
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).