Suppose I have 1Tb of data on a solaris file server. Call this the primary server.
Suppose I have another "secondary" server, and I want to zfs send the data from the primary to the secondary on a regular basis. Perhaps nightly. . Will my data stream going across the network require 1Tb every night? Or will the system somehow generate a delta from the last "zfs send" until the present time? . I think the secondary server will have a different set of snapshots from the primary server. I think there will be a new snapshot on the secondary server for every successful "zfs receive" . Right? . Will my secondary server require 1Tb of disk storage each night? Or will it be stored on disk as a delta from the last snapshot until the present? If the system is smart enough to send a delta from the last "zfs send" until the present . Then how would the system handle multiple secondary servers? Suppose I "zfs send | ssh fooserver 'zfs receive'" and suppose I "zfs send | ssh barserver 'zfs receive'" every night. Will the system be smart enough to know "I'm sending fooserver the delta from the last time I did a zfs send to fooserver until now, and I'm sending barserver the delta from the last time I did a zfs send to barserver until now." . simultaneously keeping track of both? I have to assume based on this logic, that the data stream will be 1Tb for every zfs send, and possibly the disk storage at the destination will be 1Tb for every zfs receive. Wondering if I'm wrong, and if so, how am I wrong?
_______________________________________________ opensolaris-discuss mailing list [email protected]
