(Illustration by Gaich Muramatsu)
Bradley, Bear in mind that I am still new to Coda Group, however, I would recommend a seperate partition, if that is an option for you, for /rvm. The size of this partition should be aprox. the size you specifiy to venus for your cache (I would actually recommend about 5% more space than that for a mounted fs). Also, remember that rvm under non-linux systems is usually a raw partition utilized in a manner similar to swap (but it isn't really the same thing). Our Linux setup program is designed to use a normal linux fs for the /rvm tree to make it easy to get up an running and be useful but it isn't very scalable but it does work well for small needs. But yes, if the area you are using for rvm becomes full, very bad things happen :( and we are working on ways to help ease how the system reacts when such things happen. regards, henry m. pierce ------------------------ Henry M. Pierce Research Programmer Carnegie Mellon University 5000 Forbes Avenue Pittsbugh, PA hmpierce_at_cs.cmu.edu hmp_at_infomagic.com On Tue, 21 Oct 1997, Bradley C. Kuszmaul wrote: > Date: Tue, 21 Oct 1997 13:15:26 -0400 > From: "Bradley C. Kuszmaul" <bradley_at_GRANITE.SYSTEMSX.CS.YALE.EDU> > Reply-To: linux-coda_at_TELEMANN.coda.cs.cmu.edu > To: linux-coda_at_TELEMANN.CODA.CS.CMU.EDU > Subject: when /rvm fills up /, bad things happen > Resent-Date: Tue, 21 Oct 1997 13:15:28 -0400 > Resent-From: linux-coda_at_TELEMANN.coda.cs.cmu.edu > > I thought I would drop you a note about a problem I had with coda: > > I allowed / to fill up, and /rvm is on /, which caused my system to > get very unhappy. The error messages were cryptic, and nothing in the > log seemed to help me see what was going on (it is tough to write to a > log when the log device is full though....) Venus got so confused > that one client had to be rebooted. > > By the way, CVS seems to work fine for the sandboxes if I keep the > repository off coda. (I had some trouble because my / filesystem > filled up, and I thought CVS was the problem, but now I suspect it was > the full filesystem problem.) > > -Bradley >Received on 1997-10-21 15:07:05