We help IT Professionals succeed at work.

Datastore conflicts with an existing datastore when trying to add a ESX 6.7 hosts into a VCenter cluster.

Datastore conflicts with an existing datastore when trying to add a ESX 6.7 hosts into a VCenter cluster.

This happened when I applied the lastest non-critical and critical patches to ESXi host 6.7U3. The VM's were moved to another hosts and the hosts rebooted but it was in disconnected state and when I went to try to connect it to the VCenter cluster after 15 minutes, I got the "Datastore conflicts with an existing datastore" error.

I removed the hosts from the inventory and tried again but got the same error. It did notify me of the database in conflict but the vm1 numbers for the datastore naa on all hosts is the same.

3 x ESXi Hosts 6.7 (2 are in the cluster). The third is the one I am trying to re-add.

I have the following enabled:

VSphere DRS On
VSphere HA On
VM EVC Disabled

Storage Cluster Manual Mode
VSphere Storage DRS On

What is the recommended procedure to re-add the hosts and clean up the Datastore conflict.

Thank you,
Comment
Watch Question

Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
Find the datastore in conflict, and unmount the datastore and remount.

Author

Commented:
I did that on the host but still got the same error. should I be doing the unmount and remount on all hosts or just the hosts I need to add back to the cluster.

As a last resort, could I completely reinstall ESXi on the host and rejoin the cluster?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
The datastore would need to be unmounted on all hosts. (the datastore in conflict)

Author

Commented:
ok, got it.

So here is my steps;

1. Move existing VM's from the datastore in conflict to another datastore.

2. Unmount the datastore in conflict from all hosts

3. Rejoin the host to the cluster

4. Remount the conflicting datastore to all hosts

5. Move back VM's to the conflicting datastore.


A couple of questions:

1. The VCSA has the latest update 6.7 U3b, when doing the patch updates, should i do the critical first and then the non critical or combine them together as thats what I did in the first place and all the pre-checks past but for some reason it  disconnected and never rejoined.

I am afraid this might happen to my other hosts again if I try to update like i did in the first place.

What do you recommend?

Author

Commented:
Here is what is happening:

I have 4 shared ISCSI datastores. I have unmounted the datastore which was giving me the error from all hosts after I moved the VM's to another datastore.

I then put the host into maintenance mode and rebooted the problem host. Tried to rejoin the host in VCenter and now its complaining about another datastore conflict. So unmounted that datastore as well.

So now I have 2 datastores unmounted in conflict that are inaccessible in VCenter because they are unmounted.

When I try to add the host which has now only 2 datastore, i get another error saying another datastore is in conflict.

Is this a DRS Storage cluster issue?

Should I disable DRS and not use Storage Cluster?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
You can disable Storage DRS at present, it will not hurt, it will just stop VMs moving around the Datastore Cluster.

But the datastores are the issue, but it cannot be the datastores, is it just this host which is complaining ?

This could indicate an issue with this host, is it the same host reporting the issue ?

Author

Commented:
its only this one host as I patched it and it wouldnt rejoin the cluster. VCenter has the latest update so all pre-checks passed.

I am moving the VM's off the third datastore it is complaining about and try again

Any issue if I just rebuild it from scratch and rejoin it to the cluster?

Author

Commented:
This is crazy. After moving the VM's of the 3rd datastore it complained about and trying to rejoin, it now complains about the 4th datastore.

Any issue if I just rebuild this host from scratch and rejoin it to the cluster?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
Are you Hosts and Storage certified, because this could be a driver issue after patching ?

This has occurred only on the patched host ?

Author

Commented:
I am not sure if the storage and host are certified. Its only this one host host that the updated patches have been applied.

The other 2 hosts in the cluster have not been patched it and I am afraid to do so.

So, can I rebuild this one host from scratch?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
I think I would be inclined to rollback this server to previous build.

Something has broken a driver...

Author

Commented:
how would i roll it back?

Author

Commented:
I am looking at some videos on the rollback process
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
when you start the host, at bootup there are options....

hit Shift+R to access Recovery Mode.

then select the Previous Build.

Select Y to roll back. ...

ESXi will now revert to the previous version.

if the problems, disappear, I would STICK on this build for the moment... and investigate your issues further.

Author

Commented:
ok, if I still experience the same issue, would a rebuild hurt anything especially since i have a shared datastore and drs?

Author

Commented:
The problem host build is:

[root@drtacesxihost3:~] tail -2 /*bootbank/boot.cfg
==> /altbootbank/boot.cfg <==
build=6.7.0-3.73.14320388
updated=3

==> /bootbank/boot.cfg <==
build=6.7.0-3.89.15160138
updated=4


The other 2 hosts in the cluster is:

[root@drtacesxihost1:~] tail -2 /*bootbank/boot.cfg
==> /altbootbank/boot.cfg <==
build=6.7.0-3.73.14320388
updated=3

==> /bootbank/boot.cfg <==
build=6.7.0-3.73.14320388
updated=2
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
Yes, roll back....and test.

Author

Commented:
roll back didnt work
VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017
Commented:
You will need to re-install then.

Author

Commented:
any affect on existing cluster datastores?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
No, because they are all attached to other ESXi hosts.

VMs will be stored on those datastores, and VMs will run on other hosts (compute).

The logic is, if all has gone wrong after you updates THIS host, something in the update has caused a driver issue with your storage HBA (driver).

Author

Commented:
ya man uber frustration. ok, thanks, will rebuild now
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
OR, if you have certified hosts and storage, and VMware Support, you could esclate it to VMware support to investigate.

At least you didn't upgrade them all!

Author

Commented:
no support here. will reinstall now. i did one non important host just to see, good thing
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
If you have no support, then you will have to stick at 6.7U3 and re-test this host.

Author

Commented:
got it, thanks

Author

Commented:
I reinstalled but didn't apply any patches, rejoined the cluster, all good

Author

Commented:
thanks for the quick replies and the solution
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
something broken with a HBA driver.

Author

Commented:
Ya could be, not sure how to fix that.
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
CERTIFIED EXPERT
Fellow
Expert of the Year 2017

Commented:
You'll have to compare drivers in the new versions, and re-test but difficult without support and supported servers.

Author

Commented:
got it