Help with VMWare Datastore issue

Might get more comments if you included a lot more detail of the issue. That way anyone passing by can see whether it's something they can help with.
 
So first of all, this was a little home project that I setup a number of years ago, all of which was in perfect working order until life got a little busier.

The Server, a basic wee HP ML10v2 was setup with VMware so that I could create vm's as and when I pleased. I used to dabble in computer networking and infrastructure support, but those days have long passed.

Initially, I had a single Server 2012R2 vm which I used as the backend for Plex. And with that, I had assigned a single 4TB drive as the main media store. No RAID (I know, I know).

There server has been powered down the last 3 years and upon booting up, everything was operationally sound, but unfortunately I couldn't log in to my server 2012 vm as I forgot my credentials (Again, I know :rolleyes:).

I therefore created a new, simpler Windows 10 vm in the hope that I could simply migrate the datastore across, but this is where issues began to arise. No matter what way I configured the new vm in ESXI, that is, One Hard Drive to act as c:/ drive, and a second drive, call it d:/, to act as the main storage for Plex files, I couldn't get access to the 3.5TB worth of media I had previously accumulated (:() due to 'disk capacity' errors.

After much deliberation, and probably with the fact that im a complete novice when it comes to working inside ESXI, I decided to bite the bullet, unmount the 4TB drive, reformat it, and create a new datastore. But the **** hit the fan last night. Whilst everything had been working well in that I appeared to have setup the new vm correctly, and had worked tirelessly rebuilding my media store with some 450GB worth of media, the vm became unresponsive and within ESXI, I could no longer power it on because it had ran out of space, even though there was plenty of room left on the basic 500GB c:/ drive, and loads of room on the bulk 4TB drive.

I'll drop in some screen shots below for an overview of things. Even though there is two datastore, whenever I try to assign the bulk media store to hard disk two, the vdmc file always defaults to the smaller datastore and I have a feeling this is something to do with it, even though within windows 10, I have pooled the drives in order to access the larger 4TB drive, a screenshot I'll also include.

Ive therefore gone ahead, reformatted all my hard work again, setup a new vm, pool the drives in order to create a second E:/ drive, but I dont want to begin work until I know for sure that I have access to more than 400GB. NB: When creating the media store this time, I said the drive was larger than what it is in the mad notion that I would be able to increase capacity should I come up against the dame problem. it doesn't sound right to me, but that's why im here looking help before I begin work to rebuild my media library.

 
Last edited:
So first of all, this was a little home project that I setup a number of years ago, all of which was in perfect working order until life got a little busier.

The Server, a basic wee HP ML10v2 was setup with VMware so that I could create vm's as and when I pleased. I used to dabble in computer networking and infrastructure support, but those days have long passed.

Initially, I had a single Server 2012R2 vm which I used as the backend for Plex. And with that, I had assigned a single 4TB drive as the main media store. No RAID (I know, I know).

There server has been powered down the last 3 years and upon booting up, everything was operationally sound, but unfortunately I couldn't log in to my server 2012 vm as I forgot my credentials (Again, I know :rolleyes:).

I therefore created a new, simpler Windows 10 vm in the hope that I could simply migrate the datastore across, but this is where issues began to arise. No matter what way I configured the new vm in ESXI, that is, One Hard Drive to act as c:/ drive, and a second drive, call it d:/, to act as the main storage for Plex files, I couldn't get access to the 3.5TB worth of media I had previously accumulated (:() due to 'disk capacity' errors.

After much deliberation, and probably with the fact that im a complete novice when it comes to working inside ESXI, I decided to bite the bullet, unmount the 4TB drive, reformat it, and create a new datastore. But the **** hit the fan last night. Whilst everything had been working well in that I appeared to have setup the new vm correctly, and had worked tirelessly rebuilding my media store with some 450GB worth of media, the vm became unresponsive and within ESXI, I could no longer power it on because it had ran out of space, even though there was plenty of room left on the basic 500GB c:/ drive, and loads of room on the bulk 4TB drive.

I'll drop in some screen shots below for an overview of things. Even though there is two datastore, whenever I try to assign the bulk media store to hard disk two, the vdmc file always defaults to the smaller datastore and I have a feeling this is something to do with it, even though within windows 10, I have pooled the drives in order to access the larger 4TB drive, a screenshot I'll also include.

Ive therefore gone ahead, reformatted all my hard work again, setup a new vm, pool the drives in order to create a second E:/ drive, but I dont want to begin work until I know for sure that I have access to more than 400GB. NB: When creating the media store this time, I said the drive was larger than what it is in the mad notion that I would be able to increase capacity should I come up against the dame problem. it doesn't sound right to me, but that's why im here looking help before I begin work to rebuild my media library.


OK so sounds like you're setup again.

A few things I would suggest:
- Filesystem limitations on partition table on Win10 server(?) Not sure how you have the disks formatted and partitions setup.
- Thin provisioned disks are great, but from what I can see you have allocated more than 16TB for the VM using the 4TB disk/datastore. Not sure if this is what caused issues earlier as you mentioned having ~3.5TB of media. After formatting the datastore shows as having ~3.6TB capacity so you likely ran out of space previously on the large disk.
- That last screenshot showing max size as 408GB. Made me look back again and now I've noticed you have placed both vmdisks on "Datastore1" which shows as have 408GB free space which then presents as the max size available.

Ideally you should leave some % value of space as free/unused on a datastore. As you have provisioned the VM disks as thin, you will have to keep an eye on the used capacity within the VM and what's showing on the datastore.

I would typically only use thin provisioned VM disks if you were running on a vSAN datastore. Otherwise I would suggest shrinking the 16TB disk you have assigned down to 3TB, or maybe just slightly over that to ensure you leave enough room for anything.
 
Hi Nade, thanks for taking the time to help.

No, when I created a new vm last week I created a second virtual hard disk and gave it its full capacity (4TB) but I then ran into issues - probably because it's linked to datastore1. I done 16TB this time round on a whim hoping I wouldn't run into the disk space issue again. makes no sense I know.

The biggest problem I think I have is what you have noticed, that both disks are within the same datastore, "Datastore1". For the life of me I dont know how to setup hard disk 2, the 4TB drive, to connect with the second datastore. Even when I select the second datastore when creating a new vm, it always defaults back to Datastore1 and thus, a maximum capacity of 408GB.

As for the Thin provisioning. This is also where im unsure and simply went with the settings of what was in place in the very beginning. Happy to change that if need be.
 
Last edited:
Hi Nade, thanks for taking the time to help.

No, when I created a new vm last week I created a second virtual hard disk and gave it its full capacity (4TB) but I then ran into issues - probably because it's linked to datastore1. I done 16TB this time round on a whim hoping I wouldn't run into the disk space issue again. makes no sense I know.

The biggest problem I think I have is what you have noticed, that both disks are within the same datastore, "Datastore1". For the life of me I dont know how to setup hard disk 2, the 4TB drive, to connect with the second datastore. Even when I select the second datastore when creating a new vm, it always defaults back to Datastore1 and thus, a maximum capacity of 408GB.

As for the Thin provisioning. This is also where im unsure and simply went with the settings of what was in place in the very beginning. Happy to change that if need be.
For the larger vmdisk you will need to either migrate it to the larger datastore. Or if there is nothing yet stored on the vmdisk itself, you could delete and then create a new disk.

If this is full blown ESXi (from the screenshots I'm guessing this is 6.x), then when you create/add the new disk you can expand the options and change the placement (should say something like "store with vm") and select the larger datastore. ESX will then just create datastore folder to match the VM name and create the new vmdisk in the folder. You will also see an option to switch the disk to thick provision (lazy zeroed) if you wanted to change that as well at this point.

Should be loads of documentation online with comparisons between thin and thick provision vmdisks. There's nothing wrong with using thin, but you have clearly over-allocated compared to how much actual storage space you have available. So if you stick with thin and the size as it is, then you'll run with the risk of the physical datastore running out of space at the point where the VM writes out to the disks at ~3.6TB.
 
As above you put both disks on datastore1 which then filled up when you added the media to it. You need to move the media disk to the second datastore. If it's currently empty just delete it from the VM and add it back selecting the second datastore this time
 
Back
Top Bottom