Coming hot off the heels of the last two issues, I now have only one more to address before I start applying security patches (last step before moving forward with VM creation). I'm trying to access the drive mentioned here (WD Blue 3D NAND 500GB SATA SSD) in ESXi 6.5, but can't seem to touch it at all. When I try to make a new datastore, I receive the following error:
Do I need to enable SSH and ready the SSD from command line (If so, how)? And when following the instructions given in official documentation, I can't seem to find Virtual Flash Resource Management for the life of me. Does this feature require use of vCenter by chance? I understand that this prompt seems very fragmented, so I'll put it like this:
I want to use the SSD as Virtual Flash, first and foremost. If that can't be done, how would I go about fixing the error above to make it into a VMFS datastore instead?
If you need me to pull up logs or other materials, please let me know. I've seen other posts with this error message, but I'm not sure how to fix this myself.
After sitting on the issue for a few days, I presented the issue to a few people that I know. One of them suggested re-initialising (setting the partition table for) the SSD while it was in the server, with a different OS. This was to see if there was a possible issue/bug with either the built-in storage array controller or ESXi/vCenter. I ended up initialising the SSD while it was connected to a laptop (GPT), which unfortunately destroyed the experiment. I then connected it back to the server, but also avoided sending it through the same setup as the other SAS drives in the server, detailed here. ESXi detected it, and so did vCenter. I was then able to add it to the virtual flash resource pool That was a bit of a ride, but now the info is here for others who may not have known before.
Hey TopHatProductions115,
If you want to configure a VFFS then you do not have the follow the creation of a normal VMFS and take into account that you cannot use it to store Virtual Machines. Actually the procedure is quite simple and straight to follow. Are you following the next procedure?: Set Up Virtual Flash Resource
Also i recommend you to read the considerations for it.
Thank you replying! I'd like to clarify a few points before going further:
If you want to configure a VFFS then you do not have the follow the creation of a normal VMFS and take into account that you cannot use it to store Virtual Machines.
I understand that. As I said before, I want to use the SSD as Virtual Flash, first and foremost. I will only consider the creationg of a normal VMFS if creating a VFFS doesn't work out.
That's good to hearActually the procedure is quite simple and straight to follow.
That's the same article I linked in the OP. I have tried to follow that, but was unable to find Virtual Flash Resource Management for the life of me.Are you following the next procedure?: Set Up Virtual Flash Resource
I have already done so. But thank you for the suggestion :smileygrin:Also i recommend you to read the considerations for it.
I'll be online for most of the day today.
Hey TopHatProductions115,
Are you not able to find it? Are you using the vSphere Web Client to follow the procedure right?
I was using the HTML5 client. If this feature isn't available there (as I am beginning to suspect), then I will need some help getting to the vSphere Web Client you mentioned.
Hello TopHat,
I was asking that because is not available on the HTML5 Client. For configuring VFFS you need to login to vCenter using the Flash client (vSphere Web Client). This is the URL you have to access and you can use same credentials: https://vcenter_server_fqdn/vsphere-client
For configuring VFFS you need to login to vCenter using the Flash client (vSphere Web Client).
So I do need vCenter! In that case, I'll go on and throw that onto my Windows Server 2016 VM tomorrow afternoon, after work I'll let you know how it goes :smileygrin:
Why run vCenter Server on Windows when you can deploy the VCSA?
Agree with scott28tt.
You should deploy a VCSA and if you are planning to only have one vCenter with the PSC in Embedded deployment. This recommendation is because in vSphere 7 the External PSC topology has been deprecated same as vCenter Server in Windows Server
In my case, the Windows Server VM is all but confirmed for the project I'm undertaking. This VM will take on more roles than just vCenter. It would end up being one less VM for me to manage if I put vCenter on the already-existing VM. I did read through these, though:
It looks pretty interesting, I must admit. But I also doubt that all of my hardware would be supported in any vSphere/ESXi versions beyond 6.5. By the time I decide to go for 6.7 and beyond, I'll be looking at a newer hardware stack, which would have better official support in newer ESXi releases.
Just making sure that I'm at the correct download page for the appliance:
Here is what I've been getting on most download pages:
Hello,
It makes sense, sometimes the reality is different of what is recommended on the practice but i will recommend you to create a new VCSA because in the feature if you keep upgrading your platform or whoever do it will not deal on the task of migrating the vCenter Server.
Even if the procedure is automatically done by the vCenter installer as you are migrating it is more prone of having issues during it. However the decision is on you.
And to confirm your query, yes, that is the correct link but please make sure to select vCenter 6.5 U3k which is the lastest update and contains all the security patches and bugfixes.
I'm preparing to install vCenter sometime this Labor Day weekend! Here's to hoping everything works...
I went on and took the risk, added the vCenter Appliance since I couldn't find the native Windows Server installer:
TXP-Network Does :: ESXi Test Stream! - YouTube
I'll try setting up the SSD from vCenter in a few hours...
I managed to login to the FLEX cpanel for vCenter appliance, and found the item mentioned in the documentation:
Had to also mark the storage device (naa.600508b1001c495d5e215d914cd3984f) as Flash before proceeding. That's even though ESXi saw the device as an SSD - strange. I think this may finally be solved I'll let you know how it goes in the next post...
It would appear that vCenter also can't modify the SSD for some reason. Everytime I go to touch it at all, I get consistent "Failed to retrieve data from the server" errors. Perhaps the drive isn't compatible? I'll have to look into it, since I'm not sure if VMware has this device on any support matrix...
Hey,
Glad you made it inside the Web Client. Maybe you can see at some logs in /var/log/vpxd.log when you try to add the device.
After sitting on the issue for a few days, I presented the issue to a few people that I know. One of them suggested re-initialising (setting the partition table for) the SSD while it was in the server, with a different OS. This was to see if there was a possible issue/bug with either the built-in storage array controller or ESXi/vCenter. I ended up initialising the SSD while it was connected to a laptop (GPT), which unfortunately destroyed the experiment. I then connected it back to the server, but also avoided sending it through the same setup as the other SAS drives in the server, detailed here. ESXi detected it, and so did vCenter. I was then able to add it to the virtual flash resource pool That was a bit of a ride, but now the info is here for others who may not have known before.
This is what fixed it for me.
"Run three commands through an SSH connection. From what I have seen and found this error comes from having disks that were part of different arrays and contain some data on them. When I ran the commands I was then able to connect the data stores with no issues.
ls -lha /vmfs/devices/disks/
(Verify the disk is seen. You will probably see your disk ID then :1. This is a partition on the disk. We only need to work about the main disk ID.)
partedUtil getptbl /vmfs/devices/disks/(disk ID)
(It will probably indicate that the GPT is located beyond the end of the disk.)
partedUtil setptbl /vmfs/devices/disks/(disk ID) msdos
(The output from this should be similar to msdos and the next line will be o o o o)
I hope this helps you out."
I got this answer from another person, see the post below...
This worked for me as well. Same situation - drives from a previous array had data, and had to be wiped.
hi,
I wrote this update because I've got the same problem installing an "old" 6.7U3.
It depend from partition table of disks installed.
The first way is the same of @TopHatProductio it means remove the hard drives, put on a PC, and run DISKPART utility and change in GPT or MBR partiton
The second way by is lunch a command from SSH, I found it in this link (italian link)
https://www.computermasters.it/esxi-errore-nella-creazione-datastore-cannot-change-host-config/
partedUtil mklabel /dev/disks/DISK_UUID msdos
where DISK_UUID is the UUID number of target disk
Thanks a lot!
Marco
(sorry fo my english)