We had a 3 node Vsan cluster and every thing was working fine. We had to add another node to the cluster recently. The newly added host has been configured similar to the old 3 hosts,but I am unable to migrate vms into the new server.It fails with storage related error. Also adding the new ESXi to the Vsan cluster doesnot increase the storage space available.
Isee that when I issue storage core path list command it doesnot show any naaid in the new one.
[root@NA910907-SAL:~] esxcli storage core path list |less
sas.5001438042316210-sas.0-mpx.vmhba2:C2:T2:L0
UID: sas.5001438042316210-sas.0-mpx.vmhba2:C2:T2:L0
Runtime Name: vmhba2:C2:T2:L0
Device: mpx.vmhba2:C2:T2:L0
Device Display Name: Local ATA Disk (mpx.vmhba2:C2:T2:L0)
Adapter: vmhba2
Channel: 2
Target: 2
LUN: 0
Plugin: NMP
State: active
Transport: sas
Adapter Identifier: sas.5001438042316210
Target Identifier: sas.0
Adapter Transport Details: 5001438042316210
Target Transport Details: 0
Maximum IO Size: 4194304
Whereas in the old server I see the output as follows
[root@NA910908-SAL:/tmp] esxcli storage core path list
sas.5001438038722de0-sas.0-naa.5002538c40255593
UID: sas.5001438038722de0-sas.0-naa.5002538c40255593
Runtime Name: vmhba2:C2:T2:L0
Device: naa.5002538c40255593
Device Display Name: Local ATA Disk (naa.5002538c40255593)
Adapter: vmhba2
Channel: 2
Target: 2
LUN: 0
Plugin: NMP
State: active
Transport: sas
Adapter Identifier: sas.5001438038722de0
Target Identifier: sas.0
Adapter Transport Details: 5001438038722de0
Target Transport Details: 0
Maximum IO Size: 4194304
I have checked vmkping for vsan network and i am able to reach all the ip of the remaining 3 nodes in Vsan. Is there any fix for this? I am stuck here for the past 3 days.
Hi nagaraj_avala,
Several things going on here.
First, NAA vs MPX - Some devices do not provide an NAA number, in which case ESXi generates an MPX number to represent the disk.
Has the new host been added to the vSAN cluster? If so, did you create a disk Group for the new host?
Things to look out for:
- Make sure the new disks do not have old partitions. You can check from the UI, and also delete old partitions that way.
- If disk Group has been created, run "esxcli vsan storage list" on new host. Check that the disks are in CMMDS.
- Make sure the disk are properly tagged as SSD or HD. Also seen from UI or esxcli command above.
More detail on the config may help assisting with this.
For example, did you configure the controller as passthrough or RAID0?
What is the error you get?
Is this an AF or Hybrid config?
Hi, I am Deepak.. Naga's collegue.
This is a new server, so there are no old partition.
This a All Flash config and the disk are read as SSD when i see from GUI in devices column.
We are using a P440 controller HP Gen9 blade. it is using RAID0.
esxcfg-scsidevs -a
vmhba0 be2iscsi online iscsi.vmhba0 (0000:04:00.2) Emulex Corporation Emulex OneConnect OCe14000 iSCSI UCNA
vmhba1 be2iscsi online iscsi.vmhba1 (0000:04:00.3) Emulex Corporation Emulex OneConnect OCe14000 iSCSI UCNA
vmhba2 hpsa link-n/a sas.5001438042316210 (0000:05:00.0) Hewlett Packard Enterprise Smart Array P440
vmhba32 usb-storage link-n/a usb.vmhba32 () USB
I added the new server to the existing cluster and it doesnot allow vms to be migrated to it and new vms cannot be created. So I have removed it from cluster.
Hello Deepak/Naga,
What build version of ESXi is installed on this new host?
What driver and firmware are in use on the controller?
How is the RAID0 configured? (e.g. RAID0 applied to each individual drive).
RAID0 is only supported with specific drivers/firmware and not on 6.5/6.6 for these controllers:
https://www.vmware.com/resources/compatibility/detail.php?deviceCategory=vsanio&productid=37447
What exactly do you mean by unable to migrate VMs to this host? e.g. if it is just compute resources vMotion to this VM then it doesn't even require configured vSAN storage - just clustered properly and able to access vsandatastore.
- Can you manually register a VM on this host?
Do you have vSAN partitions correctly configured on the drives on this host?
# vdq -q
# vdq -Hi
Are the disk-groups mounted and functional?
# esxcli vsan storage list | grep -i cmmds
Is the host participating in the vSAN cluster at the vSAN level (NOT the vSphere/vCenter level)?:
# esxcli vsan cluster get
Bob