This content has been marked as final. Show 1 reply
flarcreate will only include data from the root pool (typically called rpool). The odds are that this is what you actually want.
Presumably, on a 25k you would have one pool for storing the OS and perhaps home directories, etc. This is probably from some sort of a disk tray. The other pools are likely SAN-attached and are probably quite large (terabytes, perhaps). It is quite likely that instead of creating a multi-terabyte archive, you would instead want an archive of the root pool (10's to 100's of megabytes) and would use SAN features to make the other pools visible to the M9000.
One thing that you need to do that probably isn't obvious from the documentation is that you will need to add dataset resources to the solaris10 zone's configuration to make the other zpools visible in the solaris10 zone. Assuming that these other pools are on a SAN, the zpools are no longer imported on the 25k, and the SAN is configured to allow the M9000 to access the LUNs, you will do something like the following for each zpool:
In the event that you really do need to copy all of the zpools to the M9000, you can do that as well. However, I would recommend doing that using a procedure like the one described at http://docs.oracle.com/cd/E23824_01/html/821-1448/gbchx.html#gbinz. (zfs send and zfs recv can be used to send incremental streams as well. Thus, you could do the majority of the transfer ahead of time, then do an incremental transfer when you are in your cut-over window.)
# zpool import -N -o zoned=on +poolname+ # zonecfg -z +zonename+ "add dataset; set name=+poolname+; end"
If you are going the zfs send | zfs recv route and you want to consolidate the zpools into a single zpool, you can do so, then use dataset aliasing to make the zone still see the data that exists in multiple pools as though nothing changed. See http://docs.oracle.com/cd/E23824_01/html/821-1448/gayov.html#gbbst and http://docs.oracle.com/cd/E23824_01/html/821-1460/z.admin.task-11.html#z.admin.task-12.