I'm in the middle of upgrading our 3.4 box with new hardware and 4.2 Which got me thinking about putting in a cluster and possible using HA. I'm soliciting advice on how best to handle this, what I can or can't do, considering my situation.
I wanted to use two identical HP blades, one as production and the other as a mirrored backup in another building. Unfortunately having a 3rd machine is out of the budget for the moment being. The network switches in each building can only handle 1gb lines. Upgrading these is outside of my influence and just won't happen any time soon. I do have 4 lines bonded into our production, and will probably have the same for the backup. There is a NAS, but the transfer speeds to/from are slow enough to not consider running slices/virt machines on them. It does make for a great backup repository, though. All of the OS and virtual machines will be stored on a local RAID, which provides great performance for our needs.
Now for the question -
With HA and Ceph not looking like a possibility, are there any other methods that I can use to have the 2nd blade mirrored from the production? I don't mind if I have to manually turn on machines in case of a failure. I can and have tested clustering the 2 machines/nodes, but this doesn't provide mirroring as I was hoping unless I use a central drive, which is too slow in my case.
Any advice and suggestions would be greatly appreciated! If there are other discussions similar to this, please provide the link(s). Same with any tutorials.
I wanted to use two identical HP blades, one as production and the other as a mirrored backup in another building. Unfortunately having a 3rd machine is out of the budget for the moment being. The network switches in each building can only handle 1gb lines. Upgrading these is outside of my influence and just won't happen any time soon. I do have 4 lines bonded into our production, and will probably have the same for the backup. There is a NAS, but the transfer speeds to/from are slow enough to not consider running slices/virt machines on them. It does make for a great backup repository, though. All of the OS and virtual machines will be stored on a local RAID, which provides great performance for our needs.
Now for the question -
With HA and Ceph not looking like a possibility, are there any other methods that I can use to have the 2nd blade mirrored from the production? I don't mind if I have to manually turn on machines in case of a failure. I can and have tested clustering the 2 machines/nodes, but this doesn't provide mirroring as I was hoping unless I use a central drive, which is too slow in my case.
Any advice and suggestions would be greatly appreciated! If there are other discussions similar to this, please provide the link(s). Same with any tutorials.