Subscribe

FAS6210 ESXI 4.1 ISCSI Failover

Hi Netapp Forum,

Please can someone help me figure out why our esxi luns do not recover when the netapp is failed over or we lose network Ive logged calls with vmware and they are saying its the netapp and vice versa: Here is our config would like to know if there is anything wrong with it: (im just checking to make sure the netapp is setup correctly as we have ucs, vmware and netapp in this mix)

Our Netapp has 1 ip and 1 additional alias as per TR3749 the vmware host has two vmware kernels for iscsi in one vswitch with multiple tcp sessions enabled, so we see x 4 paths all active.

The vmware host is using vmware round robin psp and all 4 paths to the storage are active.

When the netapp fails over  to the other head we lose connectivity to the storage luns even though the ip and alias are available on the other head using vmkping from the esxi host, And actually even if the netapp doesnt fail over and one of the iscsi vnics fail we still lose connectivity so this looks like a esxi issue to me but need clarification....

We are using vsc 4.11, the esxi host is 4.1 U2 and the Netapp is 8.1

Questions I have is it ok for the netapp to have an ip address and an alias for redundancy using iscsi or should it just have 1 ip address bringing the paths down to 2 from 4, so each each vmkernel will only see 1 path instead of two, I thought this was the best practice.

The psp of the vmware is Round Robin but the vsc has set the SATP to VMW_SATP_DEFAULT_AA (ive seen similar non netapp bugs for this on the vmware site)

I thought the psp and satp (storage array type should be like below as the netapp igroup has ALUA enabled)

esxcli nmp satp setdefaultpsp --satp VMW_SATP_ALUA --psp VMW_PSP_RR

Please can someone provide any advice would be appreciated.

Right now Im thinking the alias on the netapp might not be supported as a valid ip path thats why it fails when there is a network failure or the SATP set by vsc is wrong?

Thanks

Re: FAS6210 ESXI 4.1 ISCSI Failover

Hi Forum:

Ive seen this link

https://communities.netapp.com/community/netapp-blogs/storage_nuts_n_bolts/blog/2009/08/26/vsphere-upgrading-from-non-alua-to-alua

As  i wrote in the first post my SATP settings is not ALUA does anyone know if this would cause the path outages when we have network issues:

As you can see my SATP is VMW_SATP_DEFAULT_AA and the esxcli nmp device list command saying it does not support path selection policy, in the link above its set to psp round robin and Storage Array Type:* VMW_SATP_ALUA*

Im wondering if I set the SATP to VMW_SATP_ALUA* the failover will work better?

naa.60a9800064664f4b6c346836766e4d65

    Device Display Name: NETAPP iSCSI Disk (naa.60a9800064664f4b6c346836766e4d65                                )

    Storage Array Type: VMW_SATP_DEFAULT_AA

    Storage Array Type Device Config: SATP VMW_SATP_DEFAULT_AA does not support                                 device configuration.

    Path Selection Policy: VMW_PSP_RR

    Path Selection Policy Device Config: {policy=rr,iops=1000,bytes=10485760,use                                ANO=0;lastPathIndex=0: NumIOsPending=0,numBytesPending=0}

    Working Paths: vmhba33:C3:T0:L12, vmhba33:C2:T0:L12, vmhba33:C1:T0:L12, vmhb                                a33:C0:T0:L12