Date: Wed, 14 Jan 2004 10:21:21 -0500 (EST) From: Robert Watson <rwatson@freebsd.org> To: Stijn Hoop <stijn@win.tue.nl> Cc: John Kennedy <jk@jk.homeunix.net> Subject: Re: CVS filling /tmp (was Re: FreeBSD CVS Question (answered!)) Message-ID: <Pine.NEB.3.96L.1040114101938.46433C-100000@fledge.watson.org> In-Reply-To: <20040114065541.GE39353@pcwin002.win.tue.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 14 Jan 2004, Stijn Hoop wrote: > On Tue, Jan 13, 2004 at 08:30:32PM -0500, Garance A Drosihn wrote: > > The problem is how CVS itself works > > with a remote repository. If you're on machine A, and the > > repository is on machine B, then CVS will build a lot of > > information in /tmp on machine B when machine 'A' does a > > cvs-checkout of the entire tree. The issue is the same if > > you do that over 'rsh' instead of 'ssh'. I do not mind the > > overhead from encrypting the connection. > > I have seen this too, and sometimes this has prevented me from checking > out RELENG_4 or some such from a machine with a swap backed /tmp. Is > there *any* workaround other than using cvsup? Is there a logical > explanation for this or is it just an unfixed bug in CVS? While I'm not quite sure of the rhyme or reason, what's occuring is that CVS is creating a mirror directory layout of the subtree you're operating on in /tmp. This means you'll actually likely run out of inodes before space, FWIW. I usually set TMPDIR to point it at /usr/tmp to avoid this being a problem. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Senior Research Scientist, McAfee Research
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1040114101938.46433C-100000>