#ontap FC with CSI driver

1 messages · Page 1 of 1 (latest)

cyan solstice
#

Hi, we are having ontap netapp with FC.
We wanted to run the CSI with kubernetes.
Is it ok to go for ontap-san driver?

heavy marten
#

sure why not? There's a minimal config example here and the steps to prepare the worker nodes here

cyan solstice
# 
version: 1
backendName: fcp-backend
storageDriverName: ontap-san
managementLIF: 10.0.0.1
svm: svm_fc
username: vsadmin
password: password
sanType: fcp
useREST: true
#

We are configuring like above

#

But we are getting the below error

#

could not configure storage pools : Could not get storage pools from array

#

above is our backend.json When we configure the backend we are getting the above erro

#

@heavy marten , pls help me to solve this issue

heavy marten
#

did you configure management on the SVM/cluster correctly? are your IP address, usename and password correct? "could not get storage pools" is usually a sign that the management connection from the worker to ONTAP fails so you should troubleshoot that

#

also you can try running Trident in debug mode for more verbose logging and check event logs in ONTAP etc

cyan solstice
#

Sure @heavy marten .. My IP username/password is correct...

#

Do i need to create the volume on top of aggregator

#

?

#

Do i need to configure the backendpools manually wiyth the volume

heavy marten
#

what IP are you using, cluster mgmt or svm mgmt? is the IP reachable from the worker nodes?

cyan solstice
#

cluster mgmt

heavy marten
#

then vsadmin is the wrong user

cyan solstice
#

IP is reachable from worker node

#

correct

#

My username is admin

#

Is it correct to give the cluster mgmt IP ?

heavy marten
#

yeah, as described here either admin is fine

cyan solstice
#

Ok

#

Do i need to create the volume ...

#

And that Should i give as backendpool

#

I am using Netapp FC

heavy marten
cyan solstice
#

Ok. Understood. Let me debug the issue why i could not create the backend..

heavy marten
#

FCP is relatively new to Trident so there might still be some issues here and there... but from the limited details you have given this looks more like a communication issue than a fundamental issue with FCP

cyan solstice
#

Ok

#

Let me check

#

Thanks for your guidance

steady cape
#

out of curiosity, do you know if some "aggregates" (or "pools") have been assigned to the ONTAP SVM ?

heavy marten
cyan solstice
#

Hi @heavy marten / @steady cape

#

Thanks for your help

#

I am able to create the PVC

#

When i am trying to attach the pod with PVC , i am getting the below error

#

failed to stage volume: no FCP hosts found for target

#

I am finding the solution for it

#

if you have any suggestions, pls share it with us

heavy marten
#

did you check the igroup on the SVM (igroup show)? Has the WWPN been added properly? what does multipath -ll show on the worker node?

#

Have you configured Zoning on your FC switches correctly?

cyan solstice
#

Thanks @heavy marten .. SOme of my configuration mismatch.. Our team have identified it based on your suggestion....

#

Is it possible to specify the thick and thin provisoning of the persistent volume claim ?

heavy marten
cyan solstice
#

Thanks @heavy marten ...

#

I want to have 2 storage class... 1 with thick and another with thin provisoning... is it good to keep 2 different backend and create the storage class individually.

heavy marten
#

you can do it in the StorageClass too, under the parameters: section, although I don't see a benefit of using thick provisioned storage

cyan solstice
#

I agree.

#

But its an requirement

cyan solstice
#

May i know which parameter will help us to enable as thick provisoning under storageClass. When i enable spaceAllocation: true under parameter section , it throwing the error

#

@heavy marten

#

pls share it when you have it

steady cape
#

=> provisioningType