This may seem like a beginner quesiton but I have never seen it before. I have a newly installed Windows 2003 Guest server that has two drives.
c:\ is 20G
d:\ is 230G
In ESX
Hard Disk 1 points to filesvr/filesvr.vmdk
Hard Disk 2 ponits to filesvr/filesvr_1-000001.vmdk
On the datastore I see
filesvr.vmdk = 20G
filesvr_1.vmdk = 240G
filesvr_1-000001.vmdk = 80G
When I look in snapshot manager all I see is you are here. No snapshots to delete or anything like that.
We are now out of space on the datastore. Any suggestions?
filesvr_1-000001.vmdk = 80G
This is the write buffer linked to the original vmdk. Looks like you took a snapshot of this at some point and never removed it, so the redo log is continually growing.
When I look in the snapshot manager for this guest all I see is "You are here" with no snapshots that can be removed. What do I do to merge this snapshot in?
well, since you are running out of space, it may not be showing you that you really are running on the snapshot. You may have to commit the changes by removing the snapshot from the command line in the service console....
vmware-cmd will allow you commit the changes and remove the snapshot, however I'm not certain on the syntax. I think it may be vmware-cmd configfile.vmx removesnapshots ...or you may be experiencing something else entirely. Nonetheless, take a look at this thread, I think it may be of use to you.
http://communities.vmware.com/thread/53295
If it doesn't help, I would open up a case with VMWare....
You also need to make sure and have enough free space to commit the snapshot. You will need a minimum amount of free space equal to the size of the snapshot(s).
OK we have successfully deleted our snapshots and merged them. When I add up the total space used I get around 400G but on our 530G datastore there is only 4G free. Where is the rest of that space at? I imagened once the snapshots were merged we would regain that space. What am I missing now?
in your VIC, open the datastore and check each of the folders...are there any files other than the vmdk's that are taking up alot of space?
No that is what is weird. When I add up all the used space I should have over 100G available.
Open up an ssh session to the ESX host. Navigate into the /vmfs/volumes/%LUN_PATH% directory, and put in the following command
du -ah --max-depth=1 | sort -n
This will determine the disk space consumed by the different folders contained on that LUN. You can go into any directory that seems to be consuming too much space and run the command again to determine where the space is being used.
The results:
3.0M ./.fbb.sf
4.0M ./.vh.sf
21G ./Gallagher-5
29G ./Gallagher-4
61M ./.fdc.sf
92G ./SQLsvr
244M ./.pbc.sf
249M ./.sbc.sf
251G ./FILEsvr
391G .
As I read these results we are using a total of 391G
Why in Vmware Infastructure Client does it report...
OK I just looked at my datastore and now it is showing the correct 146G free. So why did it not report that correctly before? It almost looks like a simple du command made the VIC start reporting the correct size.
du -ah --max-depth=1 | sort -n
sweet little command...I'll have to remember that one, thanks.