#BlueXP Disaster Recovery datastore mount issue.

1 messages · Page 1 of 1 (latest)

grim surge
#

Hey everyone! I've been trying to give the new BlueXP Disaster Recovery Free Trial a test run as a possible replacement for our SRM environment. I have my two vCenters configured, a resource group defined and a Replication Plan in place. When I attempt to run a Test Failover, it's creating the flexclone of the volume at the recovery site just fine but when it attempts to mount that new volume as a datastore within vSphere, it appears to be trying to use the cluster management IP address instead of the IP of the LIF we have defined for NFS traffic. NFS traffic is on a different network than the cluster management which is causing the mount command in vSphere to fail b/c the volume can't be mounted across the management network.

Does anyone know of a way to configure Disaster Recovery to use the proper (or different) network when mounting a datastore within vSphere?

Should've probably made mentioned that this is all on-prem; filer and vSphere.

bleak ibex
#

Not sure how much you need to preconfigure before BlueXP DR can do it's magic. So I would check if NFS is already running in your DR SVM, maybe check if the export-policy of the SVM root vol is open enough so the clone would be accessible and confirm that you have VMkernel Ports for NFS on your DR hosts.
Are the NFS IPs (from both ESXi host and SVM) in the same subnet? Or do you need routing?

grim surge
# bleak ibex Not sure how much you need to preconfigure before BlueXP DR can do it's magic. S...

All of our NFS storage traffic is on a 192.168.40.x network. The Cluster mgmt is on a 10.10.8.x network. BlueXP DR is using the 'default' export policy which is defined for all hosts in the 192.168.40.0/24. When it tries to mount the volume in vSphere, I can see that it's attempting to mount to 10.10.8.x:/volume instead of 192.168.40.x:/volume. Changing my environment to allow it to use the 10.10.8.x for mounting is not an option.

bleak ibex
#

Is it really trying to use the cluster mgmt LIF and not maybe the SVM mgmt LIF (if there is any)? Because that would be really stupid (you can't access any data volume via LIFs in the admin SVM) and most likely a bug.
Did you confirm that NFS is actually running? nfs show
I could imagine it looks for SVMs with NFS running and then for data LIFs.

#

I would suggest to create a case, they're usually quite fast (you could even try the chat).

grim surge
bleak ibex
#

hm... try the following: clone the default service-policy for NAS (default-data-files), then remove the NFS-Service and configure that policy for your CIFS-LIFs
Then do it again, remove the CIFS-Service and configure it for your NFS-LIF

You could even try to remove everything else so that only data-core and data-nfs remains, and use that service-policy for your NFS-LIFs.