From owner-freebsd-scsi@FreeBSD.ORG Wed Dec 16 07:20:12 2009 Return-Path: Delivered-To: freebsd-scsi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 094E0106566B; Wed, 16 Dec 2009 07:20:12 +0000 (UTC) (envelope-from danny@cs.huji.ac.il) Received: from kabab.cs.huji.ac.il (kabab.cs.huji.ac.il [132.65.16.84]) by mx1.freebsd.org (Postfix) with ESMTP id 895748FC0A; Wed, 16 Dec 2009 07:20:11 +0000 (UTC) Received: from pampa.cs.huji.ac.il ([132.65.80.32]) by kabab.cs.huji.ac.il with esmtp id 1NKoB6-000CQP-Ko; Wed, 16 Dec 2009 09:20:08 +0200 X-Mailer: exmh version 2.7.2 01/07/2005 with nmh-1.2 To: Miroslav Lachman <000.fbsd@quip.cz> In-reply-to: <4B281279.6060706@quip.cz> References: <4B281279.6060706@quip.cz> Comments: In-reply-to Miroslav Lachman <000.fbsd@quip.cz> message dated "Tue, 15 Dec 2009 23:49:29 +0100." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 16 Dec 2009 09:20:08 +0200 From: Daniel Braniss Message-ID: Cc: freebsd-scsi@freebsd.org, freebsd-stable Stable Subject: Re: iSCSI initiator and Dell PowerVault MD3000i X-BeenThere: freebsd-scsi@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: SCSI subsystem List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Dec 2009 07:20:12 -0000 > Hi all, > I am playing with iscsi_initiator on FreeBSD 7-STABLE and Dell > PowerVault MD3000i. This is the first time I am testing iSCSI... > > Does anyone have FreeBSD's iSCSI initiator in production / heavy load? > Or does somebody have experiences with Dell MD3000i? > > One thing is "poor performance" ~ 60 - 70MB/s depending on RAID level > used. (poor performance compared to plain SATA disk which have 110MB/s - > both tested for reading as it is our planned load - multimedia streaming > and downloads) > > > The other thing is some problem with compatibility of initiator and Dell > MD3000i. > > If I setup RAID 5 'Disk Group' consisted of 4x 1TB SATA drives (in > MD3000i) and then created for example 2 'Virtual Disks', both are > detected by iscontrol and added to /dev/ as da0 and da1, but da1 spams > log with messages like this: > > Dec 15 04:00:38 dust kernel: da0 at iscsi0 bus 0 target 0 lun 0 > Dec 15 04:00:38 dust kernel: da0: Fixed Direct > Access SCSI-5 device > Dec 15 04:00:38 dust kernel: da1 at iscsi0 bus 0 target 0 lun 1 > Dec 15 04:00:38 dust kernel: da1: Fixed Direct > Access SCSI-5 device > Dec 15 04:00:38 dust iscontrol[48576]: cam_open_btl: no passthrough > device found at 0:0:2 > Dec 15 04:00:38 dust iscontrol[48576]: cam_open_btl: no passthrough > device found at 0:0:3 > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): READ(6)/WRITE(6) not > supported, increasing minimum_cmd_size to 10. > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): READ(10). CDB: 28 0 0 0 > 0 0 0 0 1 0 > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): CAM Status: SCSI Status > Error > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): SCSI Status: Check > Condition > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): ILLEGAL REQUEST asc:94,1 > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): Vendor Specific ASC > Dec 15 04:00:39 dust kernel: (da1:iscsi0:0:0:1): Unretryable error > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): READ(10). CDB: 28 0 c > 7f df ff 0 0 1 0 > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): CAM Status: SCSI Status > Error > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): SCSI Status: Check > Condition > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): ILLEGAL REQUEST asc:94,1 > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): Vendor Specific ASC > Dec 15 04:00:40 dust kernel: (da1:iscsi0:0:0:1): Unretryable error > Dec 15 04:00:41 dust kernel: (da1:iscsi0:0:0:1): READ(10). CDB: 28 0 0 0 > 0 0 0 0 1 0 > > The message repeated many times. > > If I created more 'Virtual Disks' (7 for example), 3 of them are > producing same errors (da1, da3, da5) > > If there is only one 'Virtual Disk', it seems fine... until I configured > second path to the virtual disk as I want to try gmultipath or geom_fox > (MD3000i is dual controller with 4 NICs), then second session produces > same errors. > > First path - OK: > > Dec 15 22:47:57 dust kernel: da0 at iscsi0 bus 0 target 0 lun 0 > Dec 15 22:47:57 dust kernel: da0: Fixed Direct > Access SCSI-5 device > Dec 15 22:47:57 dust iscontrol[52226]: cam_open_btl: no passthrough > device found at 0:0:1 > Dec 15 22:47:57 dust iscontrol[52226]: cam_open_btl: no passthrough > device found at 0:0:2 > Dec 15 22:47:57 dust iscontrol[52226]: cam_open_btl: no passthrough > device found at 0:0:3 > > > Second path - error: > > Dec 15 22:48:04 dust kernel: da1 at iscsi0 bus 0 target 1 lun 0 > Dec 15 22:48:04 dust kernel: da1: Fixed Direct > Access SCSI-5 device > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): READ(6)/WRITE(6) not > supported, increasing minimum_cmd_size to 10. > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): READ(10). CDB: 28 0 0 0 > 0 0 0 0 1 0 > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): CAM Status: SCSI Status > Error > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): SCSI Status: Check > Condition > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): ILLEGAL REQUEST asc:94,1 > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): Vendor Specific ASC > Dec 15 22:48:05 dust kernel: (da1:iscsi0:0:1:0): Unretryable error > Dec 15 22:48:05 dust iscontrol[52230]: cam_open_btl: no passthrough > device found at 0:1:1 > Dec 15 22:48:05 dust iscontrol[52230]: cam_open_btl: no passthrough > device found at 0:1:2 > Dec 15 22:48:05 dust iscontrol[52230]: cam_open_btl: no passthrough > device found at 0:1:3 > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): READ(16). CDB: 88 0 0 0 > 0 1 5d 21 1f ff 0 0 0 1 0 0 > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): CAM Status: SCSI Status > Error > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): SCSI Status: Check > Condition > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): ILLEGAL REQUEST asc:94,1 > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): Vendor Specific ASC > Dec 15 22:48:06 dust kernel: (da1:iscsi0:0:1:0): Unretryable error > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): READ(10). CDB: 28 0 0 0 > 0 0 0 0 1 0 > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): CAM Status: SCSI Status > Error > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): SCSI Status: Check > Condition > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): ILLEGAL REQUEST asc:94,1 > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): Vendor Specific ASC > Dec 15 22:48:07 dust kernel: (da1:iscsi0:0:1:0): Unretryable error > Dec 15 22:48:09 dust kernel: (da1:iscsi0:0:1:0): READ(10). CDB: 28 0 0 0 > 0 1 0 0 1 0 > > > # main path > storage_1 { > initiatorname = iqn.2005-01.il.ac.huji.cs::dust.example.com > TargetName = > iqn.1984-05.com.dell:powervault.md3000i.60026b900042587b000000004ae58efc > TargetAddress = 192.168.130.101:3260,1 > tags = 64 > } > > # second path > storage_2 { > initiatorname = iqn.2005-01.il.ac.huji.cs::dust.example.com > TargetName = > iqn.1984-05.com.dell:powervault.md3000i.60026b900042587b000000004ae58efc > TargetAddress = 192.168.132.102:3260,2 > tags = 64 > } > > Can somebody advice some tweaks to get better performance and solution > of the errors above? hi Miroslav, firstly, in case you haven't yet, get the latest from: ftp://ftp.cs.huji.ac.il/users/danny/freebsd/iscsi-2.2.3.tar.gz the slowness is probably due to the scsi errors, which I need some scsi expert (hence the cc to scsi@freebsd.org, hint, hint). In the mean time, and if you can/want, you can allow me access to an iscsi partition so that I can better debug the issue. oh, and yes, we use it here. danny