Hello Jeff, Thursday, June 1, 2006, 10:36:18 AM, you wrote:
>> That helps a lot - thank you. >> I wish I knew it before... Information Roch put on his blog should be >> explained both in MAN pages and ZFS Admin Guide - as this is something >> one would not expect. >> >> It actually means raid-z is useless in many enviroments compare to >> traditional raid-5. JB> Well, it's a trade-off. With RAID-5 you pay the RAID tax on writes; JB> with RAID-Z you pay the tax on reads. I know - I only wish I new better - raid-z should be explained better in a documentation. btw: what differences there'll be between raidz1 and raidz2? I guess two checksums will be stored so one loose approximately space of two disks in a one raidz2 group. Any other things? JB> There's also another factor at play here, which is purely a matter JB> of implementation that we need to fix. With a RAID-Z setup, all JB> blocks are written in RAID-Z format -- even intent log blocks, JB> which is really stupid. If you do a lot of synchronous writes, JB> that really hurts your write bandwidth. But it's unnecessary. JB> Since we know that intent log blocks don't live for more than a JB> single transaction group (which is about five seconds), there's JB> no reason to allocate them space-efficiently. It would be far JB> better, when allocating a B-byte intent log block in an N-disk JB> RAID-Z group, to allocate B*N bytes but only write to one disk JB> (or two if you want to be paranoid). This simple change should JB> make synchronous I/O on N-way RAID-Z up to N times faster. Would be probably very useful on nfs servers. btw: just a quick thought - why not to write one block only on 2 disks (+checksum on a one disk) instead of spreading one fs block to N-1 disks? That way zfs could read many fs block at the same time in case of larger raid-z pools. ? -- Best regards, Robert mailto:[EMAIL PROTECTED] http://milek.blogspot.com _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
