Time unavailability with failover Datamover

Question

Hello everybody,

I have somes shares NFS (which use FSN) connected to the Linux server.

I would like to know the time to unavailability if the primary datamover failed ?

I saw in other post that it depends on the load of the datamover, but the minimum time what it is ?

And on the linux server, it will be necessary to umount/remount shares ?

Thank you very much.
Regards,

dynamox · Answer

i performed failover on a datamover without any file systems the other day, it took under one minute for standby datamover to completely take over. You don't need to unmount NFS exports, they will 'pause' until the NFS server is available again.

Maillon · Answer

Thank you very much for your answer It's the same for datastores (on ESX) mounted on NFS ? Regards,

dynamox · Answer

yes, just need to make sure NFS settings on ESX and VM disk time-outs are set properly. I always keep my VMs running on NFS export from Celerra/VNX during VNX File OE upgrades (which fails-over / fails back datamovers). If you search on PowerLink on Google ..there should be some best practices for deploying vSphere on NFS.

dynamox · Answer

you bet

Maillon · Answer

Thank you very much Dynamox Regards,

christopher_ime · Answer

Just to mention it, if you use the VSI USM (Virtual Storage Integrator Unified Storage Management) plug-in to provision your NFS datastores, the default settings will take care of setting the relevant ESX/ESXi Advanced Software settings for you (Net.* and TCP.*).  The other value is that it will mount the filesystem with the general best practice for instance of uncached/directed writes.

Maillon · Answer

Thank you for your reply. Regards,

VNX

Time unavailability with failover Datamover

Was this post helpful?