From owner-freebsd-questions@FreeBSD.ORG Fri Jul 2 07:33:53 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 00106106566C for ; Fri, 2 Jul 2010 07:33:52 +0000 (UTC) (envelope-from m.seaman@infracaninophile.co.uk) Received: from smtp.infracaninophile.co.uk (gate6.infracaninophile.co.uk [IPv6:2001:8b0:151:1::1]) by mx1.freebsd.org (Postfix) with ESMTP id 26EA68FC17 for ; Fri, 2 Jul 2010 07:33:51 +0000 (UTC) Received: from seedling.local (seedling.black-earth.co.uk [81.187.76.163]) (authenticated bits=0) by smtp.infracaninophile.co.uk (8.14.4/8.14.4) with ESMTP id o627XjQd089806 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Fri, 2 Jul 2010 08:33:45 +0100 (BST) (envelope-from m.seaman@infracaninophile.co.uk) X-Authentication-Warning: lucid-nonsense.infracaninophile.co.uk: Host seedling.black-earth.co.uk [81.187.76.163] claimed to be seedling.local Message-ID: <4C2D9659.3060208@infracaninophile.co.uk> Date: Fri, 02 Jul 2010 08:33:45 +0100 From: Matthew Seaman Organization: Infracaninophile User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.1.10) Gecko/20100512 Thunderbird/3.0.5 MIME-Version: 1.0 To: Ed Flecko References: <20100701212112.GA28138@gizmo.acns.msu.edu> In-Reply-To: X-Enigmail-Version: 1.0.1 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit X-Virus-Scanned: clamav-milter 0.96.1 at lucid-nonsense.infracaninophile.co.uk X-Virus-Status: Clean X-Spam-Status: No, score=1.6 required=5.0 tests=BAYES_50,DKIM_ADSP_ALL, SPF_FAIL autolearn=no version=3.3.1 X-Spam-Level: * X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on lucid-nonsense.infracaninophile.co.uk Cc: freebsd-questions@freebsd.org Subject: Re: /boot is full after running "make installkernel" on FreeBSD 8.0 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, 02 Jul 2010 07:33:53 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 01/07/2010 22:29:54, Ed Flecko wrote: > Henrik, > When I FIRST installed 8.0, I did create a separate /home partition. > When I installed the kernel and starting running out of space in / , I > thought "O.K...I'll let FreeBSD make the partition sizes IT wants to > and see if I have the same problem, and I did. > > Apparently, 512M is just, not, quite big enough so I think I'll try 1G > to give me plenty of room. Is it time for me to start advocating "one big partition" again? This may not be the consensus view, but I have found that for a quiet life and general lack of botheration it helps to create *only two* partitions on your hard drive: b: Swap -- usually 2x RAM a: Everything else Now, I've run this setup on literally hundreds of servers without problems. The usual argument against doing this is "but a run-away process might log so much that is fills your hard drive." This is true. You might also be killed by a lightning strike the next time you leave your house. Run-away logfiles are actually pretty rare, and given that 80GB would be considered a pretty small hard drive nowadays, and you can fit a standard FreeBSD install with quite a lot of extra software inside 10GB, you're likely to have sufficient empty space that you'ld get days of warning before it caused real trouble. In which case, newsyslog(8) is your friend. Cycling logs based on size and checking that every hour will avoid almost all trouble. You do monitor disk space usage on your servers don't you? Cacti is in ports and its pretty easy to set up, as are several other alternatives. Watch this list: you'll see people having trouble with too small root partitions with great regularity. I don't think I've /ever/ seen anyone ask about dealing with a process generating huge amounts of log data. Even if you do fill up the hard drive, it's not actually guaranteed disaster. FreeBSD itself will keep running just fine. So will most web applications -- although you won't get any logging. Simply delete some of the excess files, and the system will spring back to normal function. Filling the partition certainly will crash a database, but for serious RDBMS setups, I generally make an exception and put the database working files onto their own partition[*]. Nowadays too, I much prefer using ZFS -- so I have *one* zpool from which is allocated all of the space for the zdevs on the system. This is much the best of both worlds -- you get as many filesystems as you can eat, but each of them can use as much of the total available space as it needs to. Cheers, Matthew [*] As this usually involves hardware RAID10 with plenty of cache and a BBU on at least 4 x 15k RPM SAS2 drives, it would generally be on a separate partition in any case. - -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate JID: matthew@infracaninophile.co.uk Kent, CT11 9PW -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.14 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkwtllkACgkQ8Mjk52CukIyTOwCeJYhR6kY6wxmP+WlNyGF/eJte I0wAnRuULVWsjqxFAHaL1SFFTJd2sMMW =T9JF -----END PGP SIGNATURE-----