2015-06-02 03:46 AM - edited 2015-06-02 03:57 AM
Suppose I have a volume with NFS access on 7-mode filer. The plan is to transition the volume to vs1 vserver on a cDOT cluster bu using 7mtt. vs1 has lif1 created.
My question is, how to convert the user's access to the volume from 7-mode filer to "vs1' on the cDOT cluster? Is there a way on DNS to allow users who used old IP automatically be converted to the new IP (lif1 on vs1)? Or, users would have to change the IP from old one to the new one for vs1 on their end?
Thanks in advance for sharing
2015-06-02 07:29 AM - edited 2015-06-02 02:02 PM
As I understand, one way is to use a new IP for SVM, and requires to update DNS address
To prevent the situation where users are using IP address not DNS name, one could also use the same IP for SVM on 7-mode right before cutover, and without changing DNS name.
Which method should be more commonly used?
2015-06-03 11:48 AM - edited 2015-06-03 11:52 AM
Anybody please share your idea?
Configuring the data LIFs on the SVM:
I am not sure I understand above two statements. Does that mean that I am going to use the same IP for 7-mode filer and configure it for SVM on cDOT?
2015-06-03 12:16 PM
I do transition consulting for NetApp. Some answers for you inline:
>> [netappmagic] My question is, how to convert the user's access to the volume from 7-mode filer to "vs1' on the cDOT cluster? Is there a way on DNS to allow users who used old IP automatically be converted to the new IP (lif1 on vs1)? Or, users would have to change the IP from old one to the new one for vs1 on their end?
Hadrian: When performing migration of NFS/SMB volumes, you can either migrate the source IP address from the 7-mode to the cDOT SVM, or you can use a new IP address.
Generally people migrate the source IP address only if they are migrating all volumes in the same cutover window. In that scenario, NFS clients should umount the export prior to the cutover and re-mounting should work OK. The exception is if they are using custom hosts entries. You can take a look at your important NFS clients /etc/hosts file to confirm this.
If the 7-Mode IP address is not being migrated and you desire to use new IP addresses on the target, then updating DNS and advising clients to flush DNS before reconnecting is generally good enough. Since you mentioned NFS clients, you could look into any automounter configuration or /etc/fstab configuration to validate the connectivity is by DNS name or by IP address. If it is explicitly set as IP, then the client configuration will need to be updated.
>> [netappmagic] Configuring the data LIFs on the SVM:
Hadrian: This is 7MTT advising you that when you select to migrate the 7-Mode IP address to the cDOT SVM, it will pre-provision the new LIF on the SVM in a down state, and it will be enabled during the final cutover. If you are not migrating the 7-Mode IP address, if 7MTT provisions a LIF it will be immediately available.
Please let us know if this clears it up for you or not =)
2015-06-03 01:32 PM
Your messages are very helpful. I have a few follow-ups.
The section "Configuring the data LIFs on the SVM" I excerpted from "Precutover phase" in Doc "Data and Configuration Transition Guide". Understood that I could use the existing 7-mode filer IP for the target SVM, however, in most cases, we probably use a new IP for the target SVM, since we usually don't transition all volumes on 7-mode at once.
Here is what I am still confusing about. During "Data copy phase", we have initiated baseline transfer and data should be already transfered to target SVM, which means that we have already given the IP to SVM at this phase. Correct?
If correct, then why do we still need to give an IP to SVM in Precutover phase?
2015-06-03 01:57 PM - edited 2015-06-03 01:58 PM
Not necessarily. If you are have already started baseline, 7MTT will only provision the new LIFs if you checked the box in the dotted red line below, or selected New LIF as indicated in the screen shot below.
You can see the status of your network configuration decisions of that project by looking at the Networking Tab on the dashboard as indicated in the next screen shot:
If there are no entries in the Networking Tab, then you have not selected to create any LIFs or have 7MTT migrate any 7-Mode interfaces. Which is fairly common if you already have the SVM configured and working well.
Hope this helps!
2015-06-03 02:18 PM - edited 2015-06-03 02:49 PM
Sorry, I might be missing something here.
In "Data Copy phase", we have started baseline transfer, what IP we are using to do the initialization? and what is the difference between this IP and later the IP for SVM? how do I find that out and verify what these two IP's are, respectively? please show me CLI's if you could.
Thanks for patience.
2015-06-03 03:07 PM - edited 2015-06-03 03:14 PM
When migrating data from 7-mode to cDOT, we need to consider two different interfaces:
1. The Data Copy Interface, that is the source IP address which 7MTT is using to migrate data with snapmirror from source controller to destination controller. This is chosen by default in 7MTT and is a mandatory configuration.
2. The NAS Interface, also known as the transition LIF if using the 7MTT command line. These are the interfaces which your NAS/NFS clients are connecting to the storage. 7MTT reviews your source 7-mode controller interface configuration and can re-create the source interface as a new Data LIFs on the destination SVM. The Data LIF is made online during the cutover event. This is an OPTIONAL configuration in 7MTT which is not required for a 7MTT project to be successful. What is required for a successful migration is for you to understand how your NAS clients are connecting to the source 7-mode controller, and how they will be connecting to the destination storage cluster / SVM. Whether or not 7MTT manages any Data LIF provisioning is optional and up to you.
You can view the Data Copy Interface which 7MTT is using for snapmirror by viewing the 7MTT CLI command: transition show -s <session name>. You can get a list of sessions by invoking transition show without any parameter.
Note: Multipath IP address is also highlighted. If you have multiple SnapMirror interfaces or would like attempt to increase snapmirror bandwidth, you can optionally specify a SnapMirror Multipath interface but this is not so common.
You can view any NAS interfaces aka Data LIFs which 7MTT is going to manipulate by using the 7MTT CLI command: transition lif show -s <session name>
In this screenshot 7MTT session hadrianpilot_standalone did not specify any Data LIFs to be provisioned or migrated.
In 7MTT session donotmoveip_standalone, we indicated we want 7MTT to provision a new LIF on the target SVM for use by NAS clients during cutover. As mentioend earlier in the thread, 7MTT does not perform any work to update DNS or your client configuration to inform them of the new interface, that is up to you.
If you desire to understand the destination storage cluster interfaces which are used for SnapMirror replication, and that 7MTT will use to migrate the data with, you can check the storage cluster CLI like so:
If you desire to understand which LIFs are already provisioned on the destination storage cluster, check the storage cluster CLI again. These could have been provisioned by you or by 7MTT if you had selected "New LIF" in the previously indicated screen shots.
Hope this answers the question. If so please mark it as so / hit the kudos button. If not... you know where to find me =)
2015-06-04 08:34 AM
Great messages, Hadrian, Thank you!
Now to help me fully understand the process, here is the question I need to unerstand first: What is "7-Mode system FQDN" created for, if "Data-copy IP address" is used for Snapmirror between 7-mode filer and cDOT?