From owner-freebsd-questions@FreeBSD.ORG Fri Dec 31 02:19:20 2010 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 563A3106566C for ; Fri, 31 Dec 2010 02:19:20 +0000 (UTC) (envelope-from smithi@nimnet.asn.au) Received: from sola.nimnet.asn.au (paqi.nimnet.asn.au [115.70.110.159]) by mx1.freebsd.org (Postfix) with ESMTP id 3725F8FC0A for ; Fri, 31 Dec 2010 02:19:18 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by sola.nimnet.asn.au (8.14.2/8.14.2) with ESMTP id oBV2J1Ws028130; Fri, 31 Dec 2010 13:19:03 +1100 (EST) (envelope-from smithi@nimnet.asn.au) Date: Fri, 31 Dec 2010 13:19:00 +1100 (EST) From: Ian Smith To: Chris Brennan In-Reply-To: Message-ID: <20101231105353.S36121@sola.nimnet.asn.au> References: <20101229120038.3DFB0106591A@hub.freebsd.org> <20101230133126.O36121@sola.nimnet.asn.au> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: Bruce Cran , freebsd-questions@freebsd.org Subject: Re: a new hard-drive in a 2y/o laptop X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 31 Dec 2010 02:19:20 -0000 On Thu, 30 Dec 2010 11:17:48 -0500, Chris Brennan wrote: > On Thu, Dec 30, 2010 at 12:24 AM, Ian Smith wrote: > > > I don't expect this to be anything like that. Please show a) how many > > slices you allocated and how big this FreeBSD slice is and b) how you > > partitioned the FreeBSD slice into (and sizes of) / /var/ /usr [/tmp?] > > and especially swap. > > > > I wouldn't allocate any less than 1GB for your root (/) partition esp. > > if building custom kernel/s; maybe that's fixed in sysinstall for 8.2? > I cleaned out the thread, leaving only your last bit of questions here. Goodo. I'll try chopping a bit too .. > I did apparently screw up the 'dd' cmd, I retyped it correctly, below is my > (very carefully) retyped recreation of the Fixit prompt; > > [..] > Fixit# dd if=/dev/zero of=/dev/ad4 oseek-0 bs=512 count=1 Assuming that's 'oseek=0', which is the default anyway. > 1+0 records in > 1+0 records out > 512 bytes transferred ub 0.044723 secs (11448 bytes/sec) > Fixit# fdisk -Bi /dev/ad4 > ******* Working on device /dev/ad4 ******* > parameters extracted from in-core disklabel are: > cylinders=1453521 heads=16 sectors/tracks=63 (1008 blks/cyl) > > Figures below won't work with BIOS for partitions not in cyl 1 > parameters to be used for BIOS calculations are: > cylinders=1453521 heads=16 sectors/tracks=63 (1008 blks/cyl) > > Do you want to change our idea of what BIOS thinks ? [n] > [..] > > This is where I stopped, admittedly, I do not know how to use FreeBSD's > fdisk. For the sake of brevity and to move along, I'll break fdisk here and > move back to sysinstall and provide what information I can this way. Fair enough. 'what BIOS thinks' here is fine on modern disks/boxes, but the issue here is what a new(ish) user might conceive of as 'modern'! > >From sysinstalls menu, I choose 'Standard', next is the usual message about > fdisk partitioning schemes. After this, I get a 'User Confirmation Request', > which is very similar to the warning I received above. It says > > [..] > WARNING: It is safe to use a geometry of 1453521/16/63 for ad4 on computers > with modern BIOS versions. If this disk is to be uised on an old machine it > is recommended that it does not have more then 65535 cylinders, more then > 255 heads, or more then 63 sectors per track. > > Would you like to keep using the current geometry? > > > [..] > > This is where I have two choices > > Choice 1 (YES) produces the following in fdisk when choosing 'a' to use the > whole disk. > > [..] > Offset Size(ST) End Name PType Desc Subtype > Flags > 0 63 62 - 12 unused 0 > 63 1465149105 1465149167 ad4s1 8 freebsd 165 > [..] Yes, you should go with this. 'modern BIOS versions' here refers to anything later than (roughly) the mid-90s! An 'old machine' in this context - remembering sysinstall was originally written then - was one not using LBA (logical block addressing), when 8GB was a fairly big HD at least for IDE, when the 'big guys' were mostly using SCSI disks. That message is actually a lot less scary than it was until a couple of years ago, when it used to cause much more angst and regular posts, see: http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.sbin/sysinstall/disks.c.diff?r1=1.160;r2=1.161;f=h > Choice 2 (NO) produces the following in fdisk when choosing 'a' to use the > whole disk. > [..] > If you are not sure about this, please consult the Hardware Guide in the > Documentation submenu or use the {G}eometry command to change it. Remember: > You need to eneter whatever your BIOS thinks the geometry is! For IDE, it's > what you were told in the BIOS setup. For SCSI, It's the translation mode > your controller is using. Do NOT use a ''physical geometry''. > > [..] > > [..] > Offset Size(ST) End Name PType Desc Subtype > Flags > 0 63 62 - 12 unused 0 > 63 1465144002 1465144064 ad4s1 8 freebsd 165 > 1465144065 5103 1465149167 - 12 unused 0 > [..] > > Decidedly, the end result is approximately 698GB for the usable partition, > the second choice giving me a padding on both sides of the freebsd slice. You don't say what alternative geometry you entered here, if any .. but really this whole thing needs to go away. Maybe it needs some heuristic to see if it could _even possibly_ be an ancient HD needing alternative geometry? In any case, anything after 2000 is definitely 'modern'. Copying this to Bruce Cran, who's been hacking on sysinstall lately. > Moving on now, I choose the following > > Standard MBR > > Disklebel Editor > > [..] > Part Mount Size newfs > ---- ----- ---- ----- > ad4s1a / 512MB UFS2 Y > ad4s1b swap 4096MB SWAP > ad4s1d /var 4973MB UFS2+S Y > ad4s1e /tmp 512MB UFS2+S Y > ad4s1f /usr 688GB UFS2+S Y > [..] > > Decidedly not my first choice for 8.1/amd64, but I can fix that layout > later, once I know how to get the system installed correctly. Ok, I've been hunting for a commit message I noticed relatively recently and can't find just now, but I think it was to the effect that Bruce had fixed some breakage when choosing 'A' for auto-partitioning, which you indicated having chosen above. > 'Q' to quick and continue, I choose 'Minimal' then 'CD/DVD' as my > installation media. I got the usual 'Last Chance' warning and then bam, I > get > > [..] > Unable to find device node for /dev/ad4s1b in /dev! > The creation of filesystems will be aborted. > > Couldn't make filesystems properly. Aborting. > > Installation completed with some errors. You may wish to scroll through the > debugging messages on VTY1 with the scroll-lock feature. You can also choose > "No" at the next prompt and go back into the installation menus to retry > whichever operations have failed. > > [..] > > And this is where I am left. Hopefully, I've been explicit enough this time > :D Again, if I've missed something, please let me know and I shall provide > it. Indeed you have, and sorry I missed recalling this issue till now. Bruce may have something to add, but if I'm not mistaken you may just need to NOT use 'A' with your 8.1 install media, but to enter values manually. Alternatively, this may be a good time to grab an 8.2-BETA1 disc1 or memstick image where this is likely fixed, but in any case, if I had a FreeBSD slice with even half of ~700GB I'd be very much more generous with / and /tmp, and /var if you'll be using eg big databases. HTH, Ian