Date: Thu, 10 Mar 2005 22:30:11 +0300 From: Denis Shaposhnikov <dsh@vlink.ru> To: Kris Kennaway <kris@FreeBSD.ORG> Cc: Andy Hilker <ah@crypta.net> Subject: Re: the current status of nullfs, unionfs Message-ID: <878y4vxprw.fsf@neva.vlink.ru> In-Reply-To: <20050310191759.GI34206@hub.freebsd.org> (Kris Kennaway's message of "Thu, 10 Mar 2005 19:17:59 %2B0000") References: <200503091838.06322.mi%2Bmx@aldan.algebra.com> <20050310004919.GA34206@hub.freebsd.org> <87ll8vn32j.fsf@neva.vlink.ru> <20050310160852.GB1718@mail.crypta.net> <87sm33z55v.fsf@neva.vlink.ru> <20050310191759.GI34206@hub.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
>>>>> "Kris" == Kris Kennaway <kris@FreeBSD.ORG> writes: >> >> nullfs is absolutely useless for jail's because TOO slow. >> Andy> What do you mean exactly, how do you benchmark this? >> >> Just ran build of databases/mysql41-server in a jail with nullfs >> mounted /bin /sbin /lib etc over r/w nfs mounted sandbox. Kris> And compared it to what? What was your NFS setup? What Kris> measurements did you obtain? Oops, sorry. Compared to the same configuration but unionfs -r -o ro instead of nullfs. Measurements are simple: for nullfs building is s...t...r...i...n...g by string (i mean a big pause between every compiled file). And for unionfs are stringbystringbystring :) as for just ufs mounted fs's. -- DSS5-RIPE DSS-RIPN 2:550/5068@fidonet 2:550/5069@fidonet mailto:dsh@vlink.ru http://neva.vlink.ru/~dsh/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?878y4vxprw.fsf>