On 5/22/07, Scott McClanahan <scott.mcclanahan at trnswrks.com> wrote: > We have two virtual machines on ESX 3 running CentOS 4.3 and we'd like > to turn them into some sort of highly available nfs file server. > Currently, I am familiar with DRBD and RHCS but I was curious what > others might recommend in such a situation. > > I'm not sure how easily we could use RHCS in such a situation because we > don't really have the advantage of shared storage in this setup. We are > strictly limited to the two boxes. Having said that, it would be nice > to have some of the exports running on one node (tied to a VIP) and > others on the other node (tied to another VIP) if possible and manage > those services individually if needed. So basically both nodes would be > exporting select locations at any given time on this highly available > file system. That is not 100% necessary but it would be nice if the > capability is there. I would assume some sort of fencing mechanism > would need to been in place as well. > 1) If you have a high network load.. not sure if ESX is the way to go 2) I have a limited knowledge, but NFS is not a protocol that deals with 'failover' very well. You will want to look at a cluster filesystem like GFS. > This might not be the best place to ask an ESX question but would 'raw > device mapping' work in such a scenario? I don't know much about it. > > I can give more details if needed but hopefully this is enough for some > suggestions. Thanks! > > _______________________________________________ > CentOS mailing list > CentOS at centos.org > http://lists.centos.org/mailman/listinfo/centos > -- Stephen J Smoogen. -- CSIRT/Linux System Administrator How far that little candle throws his beams! So shines a good deed in a naughty world. = Shakespeare. "The Merchant of Venice"