Hi,
I would like to know if the following improvements are possible on a future
openindiana build:
1 - Increase the maximum blocksize. Right now is set at 128k but it should
be possible to increase (I would like to try 1MB as my files are usually
big).
2 - The compression stops if the compressio
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/11/10 03:07, Carlos Varela wrote:
> Hi,
>
> I would like to know if the following improvements are possible on a future
> openindiana build:
OpenIndiana is a distribution, not a kernel development group. For that
you have to bug Illumos team.
-
Date: Mon, 22 Nov 2010 21:07:32 -0500
> From: Carlos Varela
> To: openindiana-discuss@openindiana.org
> Subject: [OpenIndiana-discuss] ZFS improvements.
> Message-ID:
>
> Content-Type: text/plain; charset=ISO-8859-1
>
> Hi,
>
> I would like to know if the follo
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/11/10 13:26, Bob Hoekstra wrote:
> The next thing would be encryption. This is critical for those of us who
> take our laptops to see clients, and thus data is a little more susceptible.
> I know this is available with Sol 11 Exp, but not yet wit
If we're making a wishlist, something more immediate (and surely easy to
implement) would be a command line option for zpool create to change the
ashift value. This allows use of "Advanced Format" drives which have 4KB
internal sector size, and export them (incorrectly, for Windows XP) as 512B
bloc
taemun,
I've made some tests with WD10EARS disks (1Tb disk, 4Kb sectors reported as
512b), I've used four of them in a raidz1 zpool to keep a copy of around 20
sparse zvols I'm exporting through comstar.
To make the story short, for sparse zvols (I don't know what happens with non
sparse zvols) w
Maurilio,
For the use-case where one is using the standard (much larger) block size of
128KB, the metadata blocks will be comparatively smaller.
My problem with the ashift=9 on 4KB sectored drives is that a write of 512B
incurs an enormous hit to latency because the drive is physically reading
4K
See,
zpool on 512b sectors disks
mauri...@krumiro:~$ zpool status
pool: iscsi
state: ONLINE
status: The pool is formatted using an older on-disk format. The pool can
still be used, but some features are unavailable.
action: Upgrade the pool using 'zpool upgrade'. Once this is done, t
taemun,
as I said, it seems the same "problem" happens on filesystems (which are 128Kb
by default), where you have a lot of small files.
Maybe if you store dvd images only you don't incur in this wastage of space :)
> My problem with the ashift=9 on 4KB sectored drives is that a write of 512B
>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/11/10 17:07, Maurilio Longo wrote:
> as I said, it seems the same "problem" happens on filesystems (which are 128Kb
> by default), where you have a lot of small files.
>
> Maybe if you store dvd images only you don't incur in this wastage of sp
Jesus Cea wrote:
> The problem is evident. If you minimal allocation size is 512 bytes, per
> file you waste 256bytes.
>
> If your minimal allocation size is 4096 bytes, per file you waste 2048
> bytes.
>
> Mean values, of course.
>
Jesus,
more than that
mauri...@krumiro:~$ ls -lR /iscsi/snv_
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/11/10 18:55, Maurilio Longo wrote:
> mauri...@rsync:~$ ls -lR /nas/test/snv_143/ | wc -l
> 175184
>
> At 2Kb, mean, per file, we have 360Mb, more or less, but here we go from
> 2,73Gb to 3,81 so it is more than 1Gb, or, nearly 8Kb per file!
If
With WD EARS disks, since they do not expose the 4K size to the OS,
there is not much to be gained with their internal 4k sector size.
Performance and reliability with these is also poor. The fixed, large
TLER defaults cause constant dropouts, and I have found from experience
that these disks do
Mark,
Mark wrote:
> With WD EARS disks, since they do not expose the 4K size to the OS,
> there is not much to be gained with their internal 4k sector size.
Yes, you need a zpool command able to handle the ashift=12.
> Performance and reliability with these is also poor. The fixed, large
> TLER
14 matches
Mail list logo
