You are not logged in.
Pages: 1
http://www.phoronix.com/scan.php?page=n … &px=ODgyNA
Has anybody tried any of that? I'd like to see a PKGBUILD or something.
Offline
why do you wanna use it, it loses in every benchmark...
Offline
Do you really expect a FS with end to end checksumming and such a large amount of features to be as fast as a "basic" filesystem?
ZFS is first and foremost about data integrity and scalabilty, anything else is secondary.
Consider a mirror setup using a standard FS , if a bit of data gets screwed on drive 1 then the same bit of data gets screwed on drive 2, not so with ZFS. With the checksumming drive 2 recognizes that bit of data is corrupted and self-heals it with the correct data.
Compare creating mirrors, RAID5, LVM's etc between a normal FS and ZFS, ie RAID 5 with 1 disk for parity
zfs create tank raidz1 /dev/ada0 /dev/ada1 /dev/ada2
Snapshots.. its simple to roll back to any of these points in time.
NAME USED AVAIL REFER MOUNTPOINT
tank/home/user001@daily.6 15.0M - 861M -
tank/home/user001@daily.5 15.5M - 862M -
tank/home/user001@daily.4 13.2M - 835M -
tank/home/user001@daily.3 1.69M - 837M -
tank/home/user001@hourly.7 31.7M - 869M -
tank/home/user001@daily.2 354K - 837M -
tank/home/user001@hourly.6 1.65M - 838M -
tank/home/user001@hourly.5 480K - 837M -
/* snip */
I'd really suggest looking at all the things ZFS offers and then how fast it is all things considered, my workstation storage pool uses mirrored WD caviar blacks 1TB's and gets 250MB | 109MB read/writes, pretty fast enough.
Offline
for more reading: https://bbs.archlinux.org/viewtopic.php?id=98551
Offline
Pages: 1