[Sheepdog] recovery doubts
sonali at servercentralen.se
Tue Jan 18 11:32:39 CET 2011
I have implemented two nodes in sheepdog cluster namely kvmbox1 and kvmbox2. On each node I have one vm. Namely, vm01 on kvmbox1 and vm02 on kvmbox2. And I have set - -copies to 2 on both nodes. I want to check, If I stop sheepdog service on kvmbox1 then vm01 will be automatically recovered/redistributed on kvmbox2. Now in this situation I want to check whether recovery/redistribution of vm01 is running or not ? Is there any way to check it apart from logs..? This is my 1st concern
My 2nd concern is, consider somehow my vm01 is crashed or lost then can I recover it back fromkvmbox2..??
If yes then how to do it ? what are command to be used for recovery?
My 3rd concern is, we used command "sheep /store_dir", as per guides I thought this /store_dir , contains/stoes all objects/replicated copies. Because of which we set around 200GB space to /store_dir. But it's not. By default all objects/replicated copies are stored under /var/lib/sheepdog/obj/. If sheepdog stores all objects under /var/lib/sheepdog/obj/ then what does "sheep /store_dir" do?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the sheepdog