Date: Wed, 11 May 2005 22:49:02 -0500 From: Eric Anderson <anderson@centtech.com> To: yfyoufeng@263.net Cc: freebsd-cluster@freebsd.org Subject: Re: Clustering with Freebsd Message-ID: <4282D22E.9030204@centtech.com> In-Reply-To: <1115869431.3938.46.camel@localhost.localdomain> References: <20050429183405.M60641@bmyster.com> <20050511121115.GA2731@torch.higis.ru> <4281FF74.6060005@centtech.com> <20050511140657.GB2731@torch.higis.ru> <4282123E.6080909@centtech.com> <1115864258.3938.15.camel@localhost.localdomain> <4282CE09.6000802@centtech.com> <1115869431.3938.46.camel@localhost.localdomain>
next in thread | previous in thread | raw e-mail | index | archive | help
yf-263 wrote: > 在 2005-05-11三的 22:31 -0500，Eric Anderson写道： > >>yf-263 wrote: >> >>>在 2005-05-11三的 09:10 -0500，Eric Anderson写道： >>> >>> >>>>Dmitriy Kirhlarov wrote: >>>> >>>> >>>>>Hi Eric! >>>>> >>>>>On Wed, 11 May 2005, Eric Anderson wrote: >>>>> >>>>> >>>>> >>>>> >>>>>>Dmitriy Kirhlarov wrote: >>>>>> >>>>>> >>>>>> >>>>>>>Hi Brent! >>>>>>> >>>>>>>On Fri, 29 Apr 2005, Brent wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>>I have a webhosting business all runnning on various Freebsd i386 boxes. >>>>>>>>I was wondering if there are any good howto's on doing this ? The >>>>>>>>services id like to cluster are, >>>>>>>>apache >>>>>>>>sendmail >>>>>>>>bind 9.2.3 >>>>>>>>UW imap >>>>>>> >>>>>>> >>>>>>>replace to cyrus imapd. >>>>>>> >>>>>>>read about gmirror, carp, ggated. >>>>>> >>>>>>How would gmirror and ggated help? >>>>> >>>>> >>>>>Replicate data between two hosts, but it not very fine solution -- cached data not replicate to second node, when first node die. >>>> >>>>And it would only be for failover, not for load distribution/balancing. >>>> >>>>We really need a clustered filesystem for FreeBSD.. >>> >>> >>>Forget recite that we are also implementing it on FreeBSD ;) For the HA >>>and HP purpose. >> >>For my purposes, I need to be able to serve data on that filesystem via > > > Use our FS, you will not need NFS anymore, cause it is also a > distributed FS. And it will also reduce the cost to 1/3 ~ 1/4. I have about 1000 high-end clients using the data - mostly linux, but also HP-UX, solaris, and openbsd. Would it work on all those just as well? I put a *huge* strain on servers. >>NFS - will that be possible too? Will this code be open source? > > > Now we are considering the possibility of open source to gain a well > project life :) I'm listening!! :) >>I'd actually like to see a FreeBSD port of GFS.. > > > Before I have worked on porting Redhat GFS to Darwin, which use FreeBSD > FS infrastructure. And the porting works on CML ;) I don't know what CML is off the top of my half awake brain, but I'd like to hear more.. Eric -- ------------------------------------------------------------------------ Eric Anderson Sr. Systems Administrator Centaur Technology A lost ounce of gold may be found, a lost moment of time never. ------------------------------------------------------------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4282D22E.9030204>