From owner-freebsd-questions@FreeBSD.ORG Sat Apr 18 03:08:24 2015 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id E2DBD890 for ; Sat, 18 Apr 2015 03:08:24 +0000 (UTC) Received: from guelah.shrubbery.net (guelah.shrubbery.net [198.58.5.1]) by mx1.freebsd.org (Postfix) with ESMTP id C24898FA for ; Sat, 18 Apr 2015 03:08:24 +0000 (UTC) Received: by guelah.shrubbery.net (Postfix, from userid 7053) id 1B767548F3; Sat, 18 Apr 2015 02:52:51 +0000 (UTC) Date: Sat, 18 Apr 2015 02:52:51 +0000 From: heasley To: freebsd-questions@freebsd.org Subject: Hot swap SATA drive & /dev nodes w/ GPT Message-ID: <20150418025251.GL54669@shrubbery.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline X-PGPkey: http://www.shrubbery.net/~heas/public-key.asc X-note: live free, or die! X-homer: i just want to have a beer while i am caring. X-Claimation: an engineer needs a manager like a fish needs a bicycle X-reality: only YOU can put an end to the embarrassment that is Tom Cruise User-Agent: Mutt/1.5.23 (2014-03-12) X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 18 Apr 2015 03:08:25 -0000 Having hot swapped a SATA drive (either on an LSI 9750 or a PEXSAT34RH SATA ctlr) with a new drive that already has GPT label on it from another fbsd machine - or having accidentally deleted the wrong JBOD and recreating it, the disk or unit do not reappear in /dev/da* or /dev/gpt. Though gpart remembers part of the information. Is there some command to cause the kernel to reread the GPT? In this specific case, da2 was detached by the controller due to ECC errors. I mangled the tw_cli command to remove da2 and killed da3. I managed to get da2 reattached, but its not in /dev, though camcontrol sees it. Yesterday, I ejected a drive on a different machine and it brushed against its neighbor and they both popped out. reinserting the neighbor resulted in similar behavior. reinserting it in another slow on the controller did not help. i inserted it in a slot on a different PEXSAT34RH ctlr and it came back normally. # camcontrol devlist -v scbus0 on isci0 bus 0: <> at scbus0 target -1 lun ffffffff () scbus1 on isci0 bus 1: <> at scbus1 target -1 lun ffffffff () scbus2 on tws0 bus 0: at scbus2 target 0 lun 0 (pass0,da0) at scbus2 target 1 lun 0 (pass1,da1) at scbus2 target 2 lun 0 (pass2,da2) at scbus2 target 3 lun 0 (da3) <> at scbus2 target -1 lun ffffffff () .... da2 at tws0 bus 0 scbus2 target 2 lun 0 da2: s/n 6VP2K70V29405300E470 detached da3 at tws0 bus 0 scbus2 target 3 lun 0 da3: s/n 5N47562429405300BF0C detached tws0: INFO: (0x04: 0x0090): Unit now in active mode: unit=2 Geom name: da2 Providers: 1. Name: da2p3 Mediasize: 17179869184 (16G) Sectorsize: 512 Stripesize: 0 Stripeoffset: 2148028416 Mode: r1w1e1 Consumers: 1. Name: da2 Mediasize: 749988741120 (698G) Sectorsize: 512 Mode: r1w1e2 Geom name: da3 Providers: 1. Name: da3p3 Mediasize: 17179869184 (16G) Sectorsize: 512 Stripesize: 0 Stripeoffset: 2148028416 Mode: r1w1e1 Consumers: 1. Name: da3 Mediasize: 749988741120 (698G) Sectorsize: 512 Mode: r1w1e2