VMware Cloud Community
ReServeAdmin
Contributor
Contributor
Jump to solution

Deleting Guest VM from Host on ESXi 3.5

Hi all,

I have an ESXi 3.5 VMware host with 2 Windows 2003 server guest systems on it. About 5 months ago I ran into an issue with the host running out of disk space because I initiated some snapshots. It caused both systems to stop and while I no longer remember all the steps I took at the time, I ended up creating a 3rd guest vm and moved the 2 vmdk disk files of the 1 troubled guest vm to the new guest vm. The second guest vm I got running again without any issues. Doing this I was able to get the troubled host running again and I now want to delete the old guest vm that has been power off and not in use. I know how to delete a guest vm (right click vm and choose delete from disk) but when I start the process it gives a warning message. The part that concerns me in this case most is the part that says, "Please note if other vm's are sharing this base disk, they will no longer have access to this disk." There are currently no vmdk disks associated with this guest vm and when you browse the datastore you see no disks in the guest vm folder. The hardware configuration for the vm only shows memory and CPU's.

Before deleting this guest vm from disk I wanted to make sure there is nothing in any configuration file still associated with the troubled guest vm that may make it find the original vmdk disks (since they are still named the same but now in a different location) and delete them because at one point they were associated with the troubled guest vm. I don't need to have my new working guest vm die because somehow it found those disks and deleted them when I delete the old troubled guest vm. As long as it only looks in the folder that was created when the guest vm was originally built and deletes the files in that folder only, I should be OK since I moved the disks to the new folder, rather than just connecting to them in their original location.

Please let me know your thoughts and thanks for your help. I know I can do a backup before attempting this deletion (I have the Veeam free backup version) but one disk is 60GB and the other is 25GB so the entire vm backup takes about 5 hours.

0 Kudos
1 Solution

Accepted Solutions
weinstein5
Immortal
Immortal
Jump to solution

I would do this  -

I suppose another way to clean this up would be to browse the datastore and go to the folder for the node I want to delete and remove all files there, then delete the folder. After that I can remove the server from inventory since there is no longer any folder or files stored on the disk. It seems like this would effectively do the same thing although I don't know if there is any other housekeeping that will be missed when removing a server this way, which is why my preference was to choose the delete server from disk method.

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful

View solution in original post

0 Kudos
5 Replies
Najd
Contributor
Contributor
Jump to solution

If you remove from Inventory it leaves all of the files associated with the virtual server on the host, vmdk's etc.  If you delete from disk it will remove the node from inventory and the vmdk's etc.  If you share a vmdk with multiple virtual nodes and delete one of the virtual nodes from disk, the vmdk will be gone and none of the other virtual nodes will be able to use it, because it is deleted.


" other vm's sharing this base disk"

This pertains to the use of snapshots. If you have more than one VM using the same base disk and you delete the base disk for any one of those VMs, then the remaining disks will become inoperable. If you're not using snapshots, no problems...


But its dangerous be careful, if you are not utilizing View or Lab Manager you are safe to click the "Delete for Disk" option


Best regards hope that helped .


Najd

0 Kudos
ReServeAdmin
Contributor
Contributor
Jump to solution


Thanks for the reply Najd. It's not completely clear if that helped or answered my question. I will try to make my situation clearer in the hopes of being more confident deleting the old node will not affect other existing nodes.

I will illustrate what I currently have and see if we can come to agreement. I have a host system with a single RAID5 storage array internal to the physical host. It has a single datastore on it where all nodes are housed. Each node has its own disks, which are stored in the node folder. SERVER1 had issues so I created SERVER2 without any disks. I then moved the 2 disks which were in the SERVER1 folder to Server2's folder. After this was done I modified SERVER2's configuration, adding 2 disks and pointed to the existing disks now in the SERVER2 folder. Everything has been running fine for 4+ months. I now want to clean up and completely remove what is left of SERVER1 from the host system. Before deleting SERVER1 from disk, I wanted to make sure when I choose this option, it will not somehow figure out the 2 disks now in the SERVER2 folder (which still have the same filename they did when in the SERVER1 folder) were at one point associated with SERVER1 and try to delete them even though they now are located in SERVER2's folder and only associated with SERVER2.

I suppose another way to clean this up would be to browse the datastore and go to the folder for the node I want to delete and remove all files there, then delete the folder. After that I can remove the server from inventory since there is no longer any folder or files stored on the disk. It seems like this would effectively do the same thing although I don't know if there is any other housekeeping that will be missed when removing a server this way, which is why my preference was to choose the delete server from disk method.

I hope that clarifies my situation and someone can confirm in my scenario deleting the server from disk will not affect the vmdk disks I moved to another folder and am using with another system.

0 Kudos
weinstein5
Immortal
Immortal
Jump to solution

I would do this  -

I suppose another way to clean this up would be to browse the datastore and go to the folder for the node I want to delete and remove all files there, then delete the folder. After that I can remove the server from inventory since there is no longer any folder or files stored on the disk. It seems like this would effectively do the same thing although I don't know if there is any other housekeeping that will be missed when removing a server this way, which is why my preference was to choose the delete server from disk method.

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
0 Kudos
ReServeAdmin
Contributor
Contributor
Jump to solution

Thanks for the reply weinstein5.

Using the manual method certainly allows me to maintain more control over the process, but are you aware of anything else that may need to be done using this method to prevent future conflicts? The thing that comes to mind first is if I try to create a new vm on the same system using the same name as the deleted vm at some point in the future. Will it allow that to be done or will there be some reference to it in a system file somewhere and prevent that name from being used. This is only an example of what is a possibility, I have no idea if that is the case or if deleting it from inventory will affectively alleviate that type of concern.

0 Kudos
ReServeAdmin
Contributor
Contributor
Jump to solution

Thanks to those who replied and provided some direction. I did complete a manual delete of the files and removed the old entry from inventory and things have been fine since. This was done a while ago but I just got time to update this post so things have been good Smiley Happy

0 Kudos