Comments on: Planning Servers for Failure http://etbe.coker.com.au/2009/11/22/planning-servers-for-failure/ Linux, politics, and other interesting things Thu, 09 Feb 2012 01:09:24 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Toby http://etbe.coker.com.au/2009/11/22/planning-servers-for-failure/comment-page-1/#comment-21597 Toby Tue, 24 Nov 2009 01:54:43 +0000 http://etbe.coker.com.au/?p=1433#comment-21597 I am not sure how familiar you are with ZFS, so - I'll summarise - the crucial advantage is that ZFS uses checksums for data and metadata and performs self healing, where RAID cannot even detect corruption. http://hub.opensolaris.org/bin/view/Community+Group+zfs/whatis A half written mirror caused by a power failure, e.g. leaves RAID with a likely undetected, unfixable integrity problem. ZFS will detect and fix it, ensuring that bad data will not reach your application or operating system. By this criterion alone, it obsoletes software RAID and practically all forms of hardware RAID. I am not sure how familiar you are with ZFS, so – I’ll summarise – the crucial advantage is that ZFS uses checksums for data and metadata and performs self healing, where RAID cannot even detect corruption. http://hub.opensolaris.org/bin/view/Community+Group+zfs/whatis

A half written mirror caused by a power failure, e.g. leaves RAID with a likely undetected, unfixable integrity problem. ZFS will detect and fix it, ensuring that bad data will not reach your application or operating system.

By this criterion alone, it obsoletes software RAID and practically all forms of hardware RAID.

]]> By: etbe http://etbe.coker.com.au/2009/11/22/planning-servers-for-failure/comment-page-1/#comment-21550 etbe Sun, 22 Nov 2009 21:55:54 +0000 http://etbe.coker.com.au/?p=1433#comment-21550 Toby: I agree that RAID doesn't solve all problems. But it does solve certain classes of reasonably common problems very well. Please explain the relevance of your comment about ZFS. Toby: I agree that RAID doesn’t solve all problems. But it does solve certain classes of reasonably common problems very well.

Please explain the relevance of your comment about ZFS.

]]> By: Toby http://etbe.coker.com.au/2009/11/22/planning-servers-for-failure/comment-page-1/#comment-21547 Toby Sun, 22 Nov 2009 17:15:53 +0000 http://etbe.coker.com.au/?p=1433#comment-21547 RAID is no panacea. Compare with ZFS. RAID is no panacea. Compare with ZFS.

]]>