Fujitsu ServerView Virtual-IO Manager V3.1, Руководство пользователя

Страница: 336 / 362

13 VIOM scenarios
Via iRMC on each server blade:
As an alternative you may log on to each iRMC and set the “Power Restore
Policy” to “Always power off” for each iRMC separately, at least for all server
blades potentially running virtual addresses. We strongly recommend that
you set it for all.
This setting prevents a failed site from automatically booting the server blade
after AC power becomes available again.
Procedure for restoring the failed site
After powering-on the failed chassis, the blades will automatically execute a
so-called inventory boot and will be switched off again automatically.
As described above, the VIOM database contains the current profile assign-
ments, but the failed MMB still contains the status prior to the disaster. Syn-
chronizing the hardware with the VIOM database may be achieved by either
restoring the whole chassis, which is preferable because it is simple and
secure.
By choosing Restore on the Setup tab, you are ensuring that all chassis com-
ponents - MMB, connection blades, server blades, and IBP - are correctly
set, but this only works if all the server blades are switched off.
Alternatively, individual server blades (slots) may be restored by applying the
Restore Slot function, but keep in mind that this only configures the selected
slot and does not restore the entire IBP. As an advantage this function works
in case if certain server blades need to be kept running.
After applying the restore function, the VIOM database and the chassis hard-
ware of the restored site are in sync again and you may move applications
back to the recovered site using the server profile assignment function.
336 ServerView Virtual-IO Manager