Add nodes in unhealth cluster

[ANTONIO_SANTOS]

Hi guys,

I have a switched cluster ONTAP with a pair of FAS8080, but a node is in takeover and i can´t do giveback because IC is down. The ONTAP version is 9.1P13.

I can´t restart all environment to try reestablish the normal operations.

So, I have another pair of FAS8080 and i want to add them to this cluster. After this operation i will can move volumes to this new pair of 8080 and decomiss the old.

The question is: can i add other pair of FAS8080 to this cluster in takeover state?

Accord https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-sag/GUID-B13FF812-785E-4E62-9914-30E7A8F51A4C.html i need to have more then half of cluster nodes up to do this. 

Thanks in advance! 

[SpindleNinja]

i'm pretty sure very no.   I would open a support case asap. 

https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-sag/GUID-B13FF812-785E-4E62-9914-30E7A8F51A4C.html

[ANTONIO_SANTOS]

Thank you @SpindleNinja for you response, but this controller is expired support. For this I put this question on community.

This FAS8080 will be replaced by a FAS8300 but the arrive of this new controller will take around 2 months.

[SpindleNinja]

Got ya.    

IC?   Do you mean the Clusternet?  What caused it to go down. 

I would also raise this question with your account team and see what they can do. 

[ANTONIO_SANTOS]

@SpindleNinja yes the IC is down and i don´t know the root cause.

I replaced the MB, IOXM, IC cable and chassi but the IC remains down.

When I say IC it's about the HA function.

Yes, i will contact the account team.

Thank you!

[SpindleNinja]

Ah got ya.    I would for sure engage with support for that. 

[Mjizzini · Accepted solution]

In a rare situation due to a PCI panic from the NVRAM card, this can trigger the NVRAM card on the takeover node to go into a hung state, disabling the interconnect.

Unable to giveback after motherboard replaced on FAS62xx/FAS80xx due to no disks attached

[paul_stejskal]

You could just SnapMirror the volumes to the other 8080s. Not as seamless, but it would work.

[ANTONIO_SANTOS]

Hi @paul_stejskal 

I have already thought about this possibility and will use it if we are unable to proceed with the insertion of new nodes into the existing cluster.

Thanks for the tip!

[ANTONIO_SANTOS]

Hi @Mjizzini this link clarified my vision, because the only two itens not replaced were NVRAM and DIMM.

I will replace the NVRAM to check if IC will become online again.

Thank you!

[ANTONIO_SANTOS]

Hi @Mjizzini today I replaced the NVRAM but the IC remains down.

I´m preparing the lab to simulate insert a pair of FAS in a cluster unhealthy. If the procedure doesn´t work i will create a new cluster  to receive snapmirrors from production storage.

Anyway, I appreciate your attention.

[SpindleNinja]

How'd this go?  I am thinking it would error.