On Thu, Jul 25, 2013 at 8:49 PM, Guido Trotter <[email protected]> wrote: > On Thu, Jul 25, 2013 at 2:24 PM, harryxiyou <[email protected]> wrote: >> On Thu, Jul 25, 2013 at 7:58 PM, Guido Trotter <[email protected]> wrote: >> [...] >>> Explicitly giving a volume name for the instance is probably a bad idea. >> >> You mean we should not add "--disk-parameters param1=value1,param2=value2" >> for "gnt-instance add" command, right? >> >> >>> Can't we just "manage" the gluster pool, and allocate names as we see >>> fit, as we do for the other storage types? >>> >> >> Now, Gluster volume has to be created on the server PC manually. And then >> tell Ganeti the volume name and server hostname by a way (Now we decide >> the way "gnt-instance add --disk-parameters hostname=value1, >> volname=value2"). >> > > Well, of course this needs to change to a system that allows ganeti to > create the volumes by itself, as we do for the other storage types. > Which is unrelated of course from "managing" gluster nodes (2nd part > of the project). >
Now, you mean, we just need use the volume of Gluster, that is to say the mount operations should also be done manually. And we need manage the Gluster volume by Ganeti itself in the second part of the project, right? The first part of the project just allows me to create a Gluster disk template instance which is similar to sharedfile (Actually, I have already finished this part), right? -- Thanks Weiwei Jia (Harry Wei)
