If DoubleTake is too expensive, then you may try to allocate windows server on Linux. That means you can use iSCSI technology to store the whole of windows system on Linux platform. And, then you can use DRBD to archive your goal.
Double-take is way too expensive. It works like rsync and file system monitoring combined, doesn't it?
Thanks for this information. I do some googling and get some hints on this subject. Can you point out any specific howto's on the web? Or, can you elaborate on your suggestion? Thanks in advance.
Try to focus on "iSCSI", "Boot from SAN", "Boot iSCSI" keys googling, and you can find answers. If not, please feel free to connect me:
c r a y *at* i v a n . c o m . t w
It's free and open, just technology exchange.
Thanks for the info.
Windows的High Availability, HA)機制:
1.DFS 複寫
(1)命名空間
命名空間類似使用者輸入網芳的進入點,首先必須指定一個指定的名稱,最好是一個大家易記的名稱,而這個命名空間是會註冊到AD之上,成為一個AD的物件,因此在設定命名空間時,可以依照使用者較為習慣或是一個較易記的名稱。也就像是所有的HA架構一般,除了本身的視別名稱以外,會再給一個叢集的名稱,而這個叢集的名稱即是命名空間。
(2)檔案複寫
複寫,是自動由伺服器之間的服務去同步,而複寫是由命名空間內的同步檔案的機制,為使整個命名空間內的檔案維持一致性,因此對於該名稱區內的檔案,會由系統自動維持其一致性,而DFS是透過網芳為基礎,因此必須在兩台成員伺服器中都設定目錄共用。
2.MSCS Cluster
MSCS是由client來決定由誰來處理服務請求,所有伺服器共用一個share storage來儲存session狀態。當第一台伺服器掛了,則由第二台(被動)伺服器接手(透過HeartBeat偵測)。被動伺服器會從share storage取出session狀態,繼續未完成的工作,以達到容錯移轉的目的。
If you have SLA to meet instead of casual usage in HA requirement, you may try to leverage on the following solution : Double Take
http://www.doubletake.com/english/Pages/default.aspx
insider提到:
http://www.doubletake.com/english/Pages/default.aspx
look promising. Have you tried its Double-Take Availability http://www.doubletake.com/english/products/double-take-availability/Pages/default.aspx?