Repacker for reiser4?

Repacker for reiser4?

[Will Smith]

Is the repacker available for reiser4?  I can see code in
the kernel but no userspace tool.  The FAQ says 'not implemented'
but the reiser4 page says 'Reiser4.1 will modify the repacker'
which implies something may already be available.

I am gradually migrating my ext3 disks to reiser4, starting
at the less critical.  No problems seen so far.


I have a suggestion for the repacker:

The repacker should benchmark the disk surface and work out
which areas are fastest (bandwidth and/or track->track seek).
Then this could be combined with a usage count on each file
to gradually move the more-accessed data to faster areas
of the disk.

This would work particularly well with LVM - you could put
one logical volume on two physical volumes - one expensive,
fast and small (e.g. 74Gb RAPTOR @ 10,000rpm) and one
big, slower and cheap (e.g. any 200Gb drive).  With typical
"90:10" file access patterns, the combined disk would soon
approach the speeds of the faster disk.


Will Smith




-- 
1400 high resolution textures - http://www.mayang.com/textures
Panoramic travel photos       - http://www.willsmith.org/pano

Re: Repacker for reiser4?

[mjt]

On Sun, Sep 12, 2004 at 11:12:11PM +0800, Will Smith wrote:
>Is the repacker available for reiser4?  I can see code in

Yes, but I've heard mixed reports.

>the kernel but no userspace tool.  The FAQ says 'not implemented'

mjt@shrike:/usr/src/linux-2.6.8.1-cko6$ ls /sys/fs/reiser4/hda2/repacker/
chunk_size  count  direction  start  start_key

it works with echo 1 > /sys/fs/reiser4/hda2/repacker/start afaik, but
I haven't tried it myself yet, as my backups are a bit out of date.

>The repacker should benchmark the disk surface and work out
>which areas are fastest (bandwidth and/or track->track seek).
>Then this could be combined with a usage count on each file
>to gradually move the more-accessed data to faster areas
>of the disk.

I also talked (months ago) with the Namesys guys about having a profiler
like that. But I just had the assumption that the beginning of the disk
is fastest, not accounting for the (very cool!) LVM aspect.

I dunno if it's ever coming, they didn't have time back then.
Maybe if we gave Namesys enough money :)

-- 
mjt

Re: Repacker for reiser4?

[Will Smith]

Markus Törnqvist wrote:
>>Is the repacker available for reiser4?  I can see code in
> 
> it works with echo 1 > /sys/fs/reiser4/hda2/repacker/start afaik, but
> I haven't tried it myself yet, as my backups are a bit out of date.
> 


Thanks Markus.

As far as I can tell (can anybody confirm/deny), it's up to userspace
to manually toggle the 'direction' flag as required before starting
the repacker - the direction seems to default to 0 ('backward') on each 
mount.

It works fine for me on 600Mb of test data which I have heavily 
fragmented with a test script.

A 'diff -r' reports no differences with the original data (stored
in ext3), and I can see the tree-frag and data-frag values
falling in measurefs.ext4.

I hope to do some timings over the next few days of how fragmentation
degrades performance of both ext3 and reiser4, given the same
sequence of operations on each filesystem.  And then, whether the
repacker fixes things for reiser4.  Naturally, ext3 fragmentation
cannot be fixed :)


Will Smith

Re: Repacker for reiser4?

[mjt]

On Tue, Sep 14, 2004 at 01:42:42AM +0800, Will Smith wrote:
>
>As far as I can tell (can anybody confirm/deny), it's up to userspace
>to manually toggle the 'direction' flag as required before starting
>the repacker - the direction seems to default to 0 ('backward') on each 
>mount.

Sure seems that way...

>sequence of operations on each filesystem.  And then, whether the
>repacker fixes things for reiser4.  Naturally, ext3 fragmentation
>cannot be fixed :)

I heard a rumor that ext3's fragmentation is fixed by fsck, can anyone
deny or confirm this?

How do you measure ext3's fragmentation, btw?
(An old reiserfs user asking ;)

-- 
mjt