FAS2240, aggregation group and HA pair

[G_MARZANO]

Hello,

we have recently bought a 2240-4 unit with 12x2TB disks. There are two default aggr0 on each controller composed from 5 disks + 1 space.

I've created on aggr0 on controller 1 a new volume (vol_TEST). When I try a "cf takeover", all VIF IPs are switched to the node where I run takeover but the vol_TEST doesn't.

Is it normal because aggr0 also have root volume and all volume inside aggr0 cannot be moved between controllers?

If so, the currently configuration in our hands isn't usable, because if controller1 fails, we lost access to all volumes attached to it. Is it correct?

Hope someone can clarify me this aspect, because I think that the reseller have sold a solution does't suit our needs.

Many thanks

[scottgelb]

All volumes are there 🙂 no issue same name aggr and vol. type "partner vol status" to see partner volumes. Or type "partner" as a toggle to switch between local ad taken over node.

[G_MARZANO]

but if a controller fail, what happen to the aggr0 of failed controller?

[scottgelb]

It is taken over too. Run "partner aggr status" and you will see it

Sent from my iPhone 4S

[G_MARZANO]

Ok, now on the next three weeks, we should receive FC mezzanine card. So if a make a "cf takeover" on controller2 and then remove controller1 to install card all NFS share still working right?

I need to know if it's so, or may be to power off all systems attached to the fas and avoid data loss and downtime.

Many thanks 🙂

[scottgelb]

Correct…if you takeover the node you can pull the node that is down.

[scottgelb]

part of any install should be a takeover/giveback test before production is put on the system.  Did you have a netapp install on this?  It sounds like you went on your own without ever having done it before...and did quite a good job since it is working

[G_MARZANO]

No, I installed the system by myself and yes was the first time on NetApp storage

So, just to know if I've understood, the "partner" instance that I see when there is a takeover is real or virtual? Because the log on controller2 tells that the controller1 will do a gracefully shutdown..

Sorry for all those questions, but as I said this is my first NetApp product and I want to be sure what i'm doing before put it in production

[scottgelb]

Partner is just he toggle to see the partner during a takeover. It is a virtual instance of the down controller running as it did on the partner.

Sent from my iPhone 4S

[G_MARZANO]

ok perfect

thank you very much Scott, you were very friendly especially to reply all my stupid questions

[scottgelb]

All good questions

Sent from my iPhone 4S

[KEVIN_DAY]

Before you go live I would suggest you ensure that the Firmware is fully up to date to the latest RC release. Also I would suggest pulling the connectivity to the unit and ensure your NFS fails over in a timely manor that doesn't affect your environment. I knwo there are some tweaks on timeouts you can perform on Windows and VMware to aid this.

[scottgelb]

8.1 just went ga so good point to upgrade to the latest before or goes live.

[aborzenkov]

Personally I would rather wait for 8.1.1 or even 8.1.2 before upgrading, unless some features are desperately needed.

[scottgelb]

2240 has to run 8.1 which has been a pain point ny name at some customers since all rc code when installed. Getting the GA stamp helps there. For customers who can run 8.0.x I agree. Worth waiting for a more baked in GA release.

[KEVIN_DAY]

Never realised that 8.1 is GA now. Yes agree put this on the 2240 before you go live as there were some stability issues fixed.