ONTAP upgrade from 9.4 to 9.5

[ShahbazRiaz]

The Client wants to upgrade ONTAP version from 9.4 to 9.5, he is asking that CIFs/NFS/iSCSI will available at end-user.

Please also share complete detail for checking NFS, CIFs session detail.

Regards,

[GidonMarcus]

Hi

I want to correct that statement of @AlainTansi . SMB users that connects with SMB2 and above should sustain the session loss due to the durable handles feature. The continues availability features mentioned does not improve the resiliency for most clients, and should only be used for supported ones such SQL and Hyper-V.

The behaviour can change from one software to another. And with the way the network is built. Occasional warning/error messages displayed to the users can happen, and process that try to write in given moment can timeout.

[AlainTansi]

@GidonMarcus  the statement i made was directly from the ONTAP 9 documentation page: https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-hacg/GUID-F61468F4-1F97-4241-A370-59C7452B389C.html

I will be glad to find any reference with recent update 🙂 

[GidonMarcus]

@AlainTansi , Agree 100% it's not you. It's the docs misleading users... I'm afraid this thread is a re-run of this one regarding CA 😕 :

https://community.netapp.com/t5/ONTAP-Discussions/CIFS-NDO-and-Continuously-Available-Shares/td-p/140215

From here:

https://www.netapp.com/us/media/tr-4100.pdf - Nondisruptive Operations with SMB File SharesONTAP9.x

For CA:

"ONTAP 9 does not support persistent handles for use cases other than Microsoft Hyper-V over SMB and SQL Server over SMB".

For durable handles:

"When a LIF is migrated between nodes, client sessions become disconnected.SMB 1 is unable to prevent the lock state from being lost during these network outages, but SMB 2+clients,by using durable handles,can re establish disconnected sessions and maintain non disruptive operations. When an SMB 2 client opens a file and establishes a lock, a durable handle is created that records session details, the most important of which is the server that is associated with it. If the session is interrupted, the file remains open on the client, and the durable handle is used to reconnect the session."

------------------------------------------

@ShahbazRiaz  I don't have a working lab at the moment to test the NFS command for you. Maybe open a new thread on it if you don't find any existing ones...

I don't have other recommendations for NFS except of:

1. look for supportability in the IMT and any notes there (support.netapp.com/matrix),

2. get an upgrade advisor from ActiveIQ (https://mysupport.netapp.com/myautosupport/home.html),

3. if you wish to dig deep - look on open/fixed bugs between the two version you're upgrading to (https://mysupport.netapp.com/NOW/cgi-bin/bol/)

For SMB, there's a whole PDF TR on the second paragraph of this comment. But the main thing I can think of is to make sure your clients indeed not connected with SMB1 using the command

cifs sessions show -protocol-version  SMB1

---

[ShahbazRiaz]

Thanks for the update,

what will best practice, if we have more than 700 hundred mix SMB session?

what about NFS session?

when I tried to run below command 

"nfs connected-clients show" command it show below error.

Error: "connected-clients" is not a recognized command

Before the upgrade, if I migrate all data LIFs to 2nd node then will data loss?

[AlainTansi]

Hi @ShahbazRiaz 

For NFS sessions, you can read on this TR on page 62 https://www.netapp.com/us/media/tr-4067.pdf
For the "nfs connected-clients show" syntax, you should be running in diag mode
::> set d
::*>nfs connected-clients show

My only guest you should still get the error is if NFS is not license on the cluster.

[AlainTansi]

Hi ShahbazRiaz

Are you having an HA-pair ?
What SMB and NFS versions are you using ?

Due to the nature of the SMB protocol, all SMB sessions except for SMB 3.0 sessions connected to shares with the Continuous Availability property set, will be disruptive. SMB 1.0 and SMB 2.x sessions cannot reconnect after a takeover event. Therefore, takeover is disruptive and some data loss could occur.

For details about CIFS and NFS sessions use the following commands
::> vserver cifs session show
::> vserver nfs show

You can also check the session uptime by running the following;
::> system node run -node <node_name> -command uptime

You can read on this articles for more details;

What happens during takeover: https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-hacg/GUID-F61468F4-1F97-4241-A370-59C7452B389C.html

Upgrading the first node in an HA pair: https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-ug-rdg/GUID-99A0D279-B5DF-4B4B-AC6E-7C3832D2C3CA.html

Please if this post resolved your issue, please help others by selecting ACCEPT AS SOLUTION or adding a KUDO or both!