Comments on: Oracle 11gR2 ASM / ACFS: A first benchmark (poorly) http://www.usn-it.de/index.php/2009/09/15/oracle-11gr2-asm-acfs-a-first-benchmark-poorly/ Martin Klier - Linux, Oracle, Asterisk Tue, 31 Aug 2010 15:51:35 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: usn http://www.usn-it.de/index.php/2009/09/15/oracle-11gr2-asm-acfs-a-first-benchmark-poorly/comment-page-1/#comment-11718 usn Wed, 16 Sep 2009 18:24:32 +0000 http://www.usn-it.de/?p=306#comment-11718 Easy to read above: "Building disk group DATA from them, with redundancy NORMAL" How ASM redundancy works isn't the topic, it does not affect the comparision between Database ASM usage and ACFS if I want to compare them as-is. The entry above describes, how I did change the mirroring within NORMAL redundancy - sorry if this wasn't clear? Regards Usn Easy to read above:
“Building disk group DATA from them, with redundancy NORMAL”

How ASM redundancy works isn’t the topic, it does not affect the comparision between Database ASM usage and ACFS if I want to compare them as-is.
The entry above describes, how I did change the mirroring within NORMAL redundancy – sorry if this wasn’t clear?

Regards
Usn

]]> By: Anonymous http://www.usn-it.de/index.php/2009/09/15/oracle-11gr2-asm-acfs-a-first-benchmark-poorly/comment-page-1/#comment-11701 Anonymous Wed, 16 Sep 2009 03:40:26 +0000 http://www.usn-it.de/?p=306#comment-11701 What is your volume redundancy? If your volume redundancy is NORMAL than every write you are doing to the disk is going to be mirrored to 2 disks. If your underlying VM storage is a single disk, then every mirrored write is doing 2 writes to the same disk. That would account for the throughput being halved. If you had 2 real disks then the mirrored writes would be issued in parallel. What is your volume redundancy? If your volume redundancy is NORMAL than every write you are doing to the disk is going to be mirrored to 2 disks. If your underlying VM storage is a single disk, then every mirrored write is doing 2 writes to the same disk. That would account for the throughput being halved. If you had 2 real disks then the mirrored writes would be issued in parallel.

]]> By: Anonymous http://www.usn-it.de/index.php/2009/09/15/oracle-11gr2-asm-acfs-a-first-benchmark-poorly/comment-page-1/#comment-11690 Anonymous Tue, 15 Sep 2009 19:10:38 +0000 http://www.usn-it.de/?p=306#comment-11690 [...] nach komplett Ende. Ich habe das Ganze heute mal getestet (wen es interessiert der kann hier: http://www.usn-it.de/index.php/2009/...chmark-poorly/ nachlesen). Mein Eindruck hinsichtlich Administration war gut, auch das Sperrverhalten scheint [...] [...] nach komplett Ende. Ich habe das Ganze heute mal getestet (wen es interessiert der kann hier: http://www.usn-it.de/index.php/2009/…chmark-poorly/ nachlesen). Mein Eindruck hinsichtlich Administration war gut, auch das Sperrverhalten scheint [...]

]]>