Re: we are finding that parity writes are half of all writes when writing 50mb files

We gave Neil Brown's patch a try, it is about three times faster than the stuff in the kernel. It avoids the unnecessary parity writes I first complained about when writing 50mb files, and we are going to document for reiserfs users that they should be using his patch. Hans - To unsubscribe fr

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Tuesday November 28, [EMAIL PROTECTED] wrote: > On Tue, Nov 28, 2000 at 10:50:06AM +1100, Neil Brown wrote: > >However, there is only one "unplug-all-devices"(*) call in the API > >that a reader or write can make. It is not possible to unplug a > >particular device, or better still

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Tue, Nov 28, 2000 at 11:46:10AM -0700, Hans Reiser wrote: > We need vm to push the FS which pushes raid, [..] This just happens. > [..] and the FS needs to know about > or even contain the stripe cache. That's not a blkdev layer issue but a FS<->LVM/RAID issue. What's suboptimal at the mome

Re: we are finding that parity writes are half of all writes when writing 50mb files

Andrea Arcangeli wrote: > > On Tue, Nov 28, 2000 at 10:50:06AM +1100, Neil Brown wrote: > >However, there is only one "unplug-all-devices"(*) call in the API > >that a reader or write can make. It is not possible to unplug a > >particular device, or better still, to unplug a particul

Re: we are finding that parity writes are half of all writes when writing 50mb files

Neil Brown wrote: > > On Monday November 27, [EMAIL PROTECTED] wrote: > > using reiserfs over raid5 with 5 disks. This is unnecessarily suboptimal, it >should be that parity > > writes are 20% of the disk bandwidth. Comments? > > > > Is there a known reason why reiserfs over raid5 is way worse

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Tue, Nov 28, 2000 at 10:50:06AM +1100, Neil Brown wrote: >However, there is only one "unplug-all-devices"(*) call in the API >that a reader or write can make. It is not possible to unplug a >particular device, or better still, to unplug a particular request. This isn't totally tru

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Monday November 27, [EMAIL PROTECTED] wrote: > using reiserfs over raid5 with 5 disks. This is unnecessarily suboptimal, it should >be that parity > writes are 20% of the disk bandwidth. Comments? > > Is there a known reason why reiserfs over raid5 is way worse than ext2. Does ext2 >optim

Re: we are finding that parity writes are half of all writes when writing 50mb files

Hans Reiser <[EMAIL PROTECTED]> writes: > Thank you, this may have been extremely valuable. Now the next > question is, wouldn't it be a lot easier to just fix the raid code > than to fix reiserfs in this area? > > Why is there only a single open stripe? Should we keep all raid > stripes open u

Re: we are finding that parity writes are half of all writes when writing 50mb files

"Harald Nordgård-Hansen" wrote: > > Hans Reiser <[EMAIL PROTECTED]> writes: > > > using reiserfs over raid5 with 5 disks. This is unnecessarily > > suboptimal, it should be that parity writes are 20% of the disk > > bandwidth. Comments? > > > > Is there a known reason why reiserfs over raid5

Re: we are finding that parity writes are half of all writes when writing 50mb files

Hans Reiser <[EMAIL PROTECTED]> writes: > using reiserfs over raid5 with 5 disks. This is unnecessarily > suboptimal, it should be that parity writes are 20% of the disk > bandwidth. Comments? > > Is there a known reason why reiserfs over raid5 is way worse than > ext2. Does ext2 optimize f

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Mon, Nov 27, 2000 at 04:58:12PM -0200, Carlos Carvalho wrote: > Jakob Østergaard ([EMAIL PROTECTED]) wrote on 27 November 2000 18:18: > >On Mon, Nov 27, 2000 at 02:24:03PM -0700, Hans Reiser wrote: > >> using reiserfs over raid5 with 5 disks. This is unnecessarily > >> suboptimal, it should

Re: we are finding that parity writes are half of all writes when writing 50mb files

Jakob Østergaard ([EMAIL PROTECTED]) wrote on 27 November 2000 18:18: >On Mon, Nov 27, 2000 at 02:24:03PM -0700, Hans Reiser wrote: >> using reiserfs over raid5 with 5 disks. This is unnecessarily >> suboptimal, it should be that parity writes are 20% of the disk >> bandwidth. Comments? > >I

Re: we are finding that parity writes are half of all writes when writing 50mb files

On Mon, Nov 27, 2000 at 02:24:03PM -0700, Hans Reiser wrote: > using reiserfs over raid5 with 5 disks. This is unnecessarily suboptimal, it should >be that parity > writes are 20% of the disk bandwidth. Comments? It would be great to have a number describing ``unnecessarily suboptimal'' more a

we are finding that parity writes are half of all writes when writing 50mb files

using reiserfs over raid5 with 5 disks. This is unnecessarily suboptimal, it should be that parity writes are 20% of the disk bandwidth. Comments? Is there a known reason why reiserfs over raid5 is way worse than ext2. Does ext2 optimize for raid5 in some way? Hans - To unsubscribe from thi