Best way to reboot a VNX SP

Why do you need to reboot SPB  ?

Using GUI is the easiser way to do, i have not seen a naviseccli option. LUNs will trespass automatically when SP is going down and will trespass back once SP is back online. If LUNs did not trespass back you can manually do with with command: naviseccli -h SPB trespass mine

Hi

There is an issue with the standby power supply wiring and the fix is to reboot the SP. We need to have a clear environment to collect health information for an upcoming block code update therefore it needs to be fixed.

Okey then i will try the GUI one and tresspass the LUNS over after it's up again because i realized last time i did a shutdown it didn't tresspass back?!

Hello Duke,

I don't see a problem in using the GUI option. For absolute precaution, you can disable the write cache (to avoid cache loss problem if the other SP should fail while this SP is rebooting. This can also be done in the GUI of Unisphere under System > manage cache.

For the CLI option, if you wish to choose it, there is this command:

naviseccli -h -user -password -scope 0 rebootsp

You can also run a different command from the other SP:

naviseccli -h -user -password -scope 0 rebootpeersp

The LUNs should trespass automatically from the SP (once its down), and trespass back automatically as well once it has booted. If any issue occurs, or if you want to manually make sure each SP grabs its own (default-owner) LUNs, you can run this command from each SP:

naviseccli -h -user -password -scope 0 trespass mine

Alternatively, you can also do this from Unisphere > Storage > LUNs and choose the Trespassed LUNs view.

Please note that SP reboot operations can be scheduled with EMC Support to be handled by either to local or Remote Change Management teams.

Hope this helps,

Adham

Hi guys

I've successfully rebooted the SP, thanks. But there is one last question I've got.

During this operation I/O to this SP should not be available because the only message I've got is:

SP A Device N/A Description Unisphere can no longer manage (SP B). This does not impact server I/O to the storage system. See alerts for details.

Is this kind of reboot just a management server reboot or a real reboot where the whole SP is going down and start again?

Thanks

What that means is the IO is not interrupted because the LUNs were trespassed to the remaining SP. In the Event log for each SP there should be a message about the SP rebooting.

glen