From owner-freebsd-stable Tue Jan 4 11:22:45 2000 Delivered-To: freebsd-stable@freebsd.org Received: from boromir.vpop.net (dns1.vpop.net [206.117.147.2]) by hub.freebsd.org (Postfix) with ESMTP id CC57D14C59 for ; Tue, 4 Jan 2000 11:22:32 -0800 (PST) (envelope-from mreimer@vpop.net) Received: from vpop.net (bilbo.vpop.net [216.160.82.65]) by boromir.vpop.net (8.9.1/8.9.1) with ESMTP id LAA10086 for ; Tue, 4 Jan 2000 11:22:30 -0800 (PST) Message-ID: <38724886.9A2DFB05@vpop.net> Date: Tue, 04 Jan 2000 11:22:46 -0800 From: Matthew Reimer Organization: VPOP Technologies, Inc. X-Mailer: Mozilla 4.7 [en] (X11; I; FreeBSD 3.4-STABLE i386) X-Accept-Language: en MIME-Version: 1.0 To: freebsd-stable@freebsd.org Subject: Unexpected vinum behavior while reviving Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG FreeBSD 3.3 circa 1999 Nov 10, two 36G disks mirrored (one plex per disk, da1 and da2). While testing vinum's ability to revive a plex, I was surprised by the pattern of disk activity. Nothing was running on the box, except vinum, reviving a 35G mirrored plex I had earlier disconnected. What seems strange is the traffic pattern between the two disks: da1 was (presumably reading) at ~2MB/s, while da2 was (presumably writing) at 6.5MB/s. Shouldn't the traffic be the same for both? Reading/writing the volume when it's up (i.e., not reviving) yields symmetrical traffic patterns. Here's systat -vmstat output: 3 users Load 0.01 0.00 0.00 Wed Nov 10 18:38 Mem:KB REAL VIRTUAL VN PAGER SWAP PAGER Tot Share Tot Share Free in out in out Act 1736 848 2780 916 222708 count All 32784 1452 2259116 1728 pages cow Interrupts Proc:r p d s w Csw Trp Sys Int Sof Flt zfod 369 total 2 2 143 90 368 71 2 15248 wire 100 clk0 irq0 5792 act 128 rtc0 irq8 0.5%Sys 0.0%Intr 0.0%User 0.0%Nice 99.5%Idl 11704 inact 140 pci irq11 | | | | | | | | | | cache 1 pci irq10 222708 free fdc0 irq6 daefr atkbd0 irq Namei Name-cache Dir-cache prcfr Calls hits % hits % react pdwake pdpgs Discs da0 da1 da2 fd0 pass0 pass1 pass2 intrn KB/t 0.00 64.00 64.00 0.00 0.00 0.00 0.00 8385 buf tps 0 35 105 0 0 0 0 dirtybuf MB/s 0.00 2.18 6.56 0.00 0.00 0.00 0.00 17268 desiredvnodes % busy 0 4 95 0 0 0 0 340 numvnodes ---------------------------- Vinum is configured like this: vinum -> l Configuration summary Drives: 2 (4 configured) Volumes: 1 (4 configured) Plexes: 2 (8 configured) Subdisks: 2 (16 configured) D vinumdrive0 State: up Device /dev/da1s1e Avail: 0/35000 M B (0%) D vinumdrive1 State: up Device /dev/da2e Avail: 0/35000 M B (0%) V fatmirror State: up Plexes: 2 Size: 34 GB P fatmirror.p0 C State: up Subdisks: 1 Size: 34 GB P fatmirror.p1 C State: faulty Subdisks: 1 Size: 34 GB S fatmirror.p0.s0 State: up PO: 0 B Size: 34 GB S fatmirror.p1.s0 State: reviving PO: 0 B Size: 34 GB After fatmirror.p1 has been revived, it looks like this: vinum -> l Configuration summary Drives: 2 (4 configured) Volumes: 1 (4 configured) Plexes: 2 (8 configured) Subdisks: 2 (16 configured) D vinumdrive0 State: up Device /dev/da1s1e Avail: 0/35000 M B (0%) D vinumdrive1 State: up Device /dev/da2s1e Avail: 35000/350 00 MB (100%) V fatmirror State: up Plexes: 2 Size: 34 GB P fatmirror.p0 C State: up Subdisks: 1 Size: 34 GB P fatmirror.p1 C State: up Subdisks: 1 Size: 34 GB S fatmirror.p0.s0 State: up PO: 0 B Size: 34 GB S fatmirror.p1.s0 State: up PO: 0 B Size: 34 GB Is this unequal traffic across the disks a bug? Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message