Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
nathan at robotics.net wrote: > On Wed, 16 Jul 2008, Bob Goldberg wrote: > >> SO - here's what I really want: 2 server's fully redundant with each >> other - >> running several vmware virtual servers. If 1 system fails, the other >> system >> seemlessly provides access to the virtual servers, without any >> interruption >> of service to the network. I would prefer a solution which only >> requires 2 >> physical servers. > > Ok, can be done. > >> Here's my current solution: drbd with HA (high availability) >> What I'm finding, is that HA is not an "instant" availability >> solution. So >> now I'm wondering if I need some other clustering solution. >> >> Anyone have suggestions on what's the best solution, and which >> applications >> provide same? > > I can't tell you if it is the best solution or not, but I have the > above working. I am running gfs2 ontop of lvm2-cluster on drbd in > primary / primary mode, HA is done with cman and IPMI fencing. > > P.S. I am using 10 gig infiniband for DRBD to keep all that traffic > off the ethernet interfaces. If you look on ebay you can buy it for > next to nothing, I paid about $45 for the 10 gig cards, $35 for cables > and $400 for a switch. > You have dual nic infiniband? Or single nic? My question it's because of the backup, i usually do bonding of two ethernet NIC's, so i can use the advantage of bonding to survive to NIC failures, my question about infiniband it's, if i buy dual nics infiniband, it's possible to do the same? Or even better, the own infiniband support this failure in hardware? Sorry for the questions, but i never worked out with infiniband. Cheers. -- Igor Neves <igor.neves at 3gnt.net> 3GNTW - Tecnologias de Informação, Lda SIP/JID/MSN: igor at 3gnt.net Mobile Phone: 00351914503611 Local Phone: 00351252377120