Learn how to a build a cloud-first strategyRegister Now

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 783
  • Last Modified:

VMware RDM lun out of space?

EE,

I'm having a hard time determining where my out of space issues on my RDM file server lun are coming from. Here are the results of df -h on my storage server.

RDM_vol = 7TB Total 7TB used
RDM_lun = 5TB occupied
File Server windows OS shows 1TB on that RDM attached lun

The mapping resides on a VMFS datastore with 50% free space. Everything is thin provisioned

So does the problem have to do with space reclamation? I didn't think that would be effective because fileservers always grow and never shrink thus the size should reflect the real size. Is there an alternative other then installing snapdrive?

Thanks again in advance
0
snyderkv
Asked:
snyderkv
  • 9
  • 6
1 Solution
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
The fact you mention Snapdrive, I assume the SAN is a NetApp ?

Is the LUN 5TB ?

How is this connected to the VM?

directly via Software iSCSI in the VM or via RDM through vSphere ESXi ?

So the size of RAW disk presented to VM, is also 5TB ?
0
 
snyderkvAuthor Commented:
Yes the RDM is attached via vSphere ESXi 5.5 and the windows operating system obviously sees it as a 5TB lun as it is directly attached to the guest VM. The RDM is also in physical mode but that doesn't matter.

The OS sees 1TB but the vol status -v RDM_lun shows 5TB occupied. I don't get it. Any other caveats? Otherwise, I'm going to snapdrive and reclaim space to see if that helps. Looks like VSC only works with NFS datastores so I can't use that.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
If the NTFS 5TB drive has been formatted as 5TB, it will occupy 5TB.

I would attach directly via Microsoft iSCSI initiator, and use Snapdrive in the VM.
0
Transaction-level recovery for Oracle database

Veeam Explore for Oracle delivers low RTOs and RPOs with agentless transaction log backup and transaction-level recovery of Oracle databases. You can restore the database to a precise point in time, even to a specific transaction.

 
snyderkvAuthor Commented:
Even if it's thin provisioned?
0
 
snyderkvAuthor Commented:
Even so, you can see the RDM volume is 7TB. 5TB should not offline the volume. Where is the extra 2TB going? The more time passes the more I forget about the situation so I might just have to close the thread.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
If the volume is going offline, it's because it's volume is full.

So any snapshots being performed on the volume?

even by default? snapshots can occur ?

Snapshot space occupies the volume

snapshot space + LUN space = Total Volume
0
 
snyderkvAuthor Commented:
200g of snapshots and snapshot autodelete is enabled with defaults.

df -h reports 7TB used all it has in it is the 5TB lun plus 200gb snapshot. I may have missed something. I fixed it too quickly by increasing the size of the volume unfortunately so now I just can't see anything unsuaul
0
 
snyderkvAuthor Commented:
I'm wondering if the volume offlined for some other reason now. Anyways closing thread
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
We have our volumes set to Autogrow to prevent LUNs going OFFLINE.

Back in the LAB tomorrow, so I'll try some examples, with 7TB volume, and 5TB LUN, RDM to VM.
0
 
snyderkvAuthor Commented:
Ok let me know of any interesting findings but don't sweet it. I may not have all the informaiton needed to provide you a starting point for troubleshooting. All I know for now is that I have 9TB volume that says 7TB used. The LUN shows 5TB occupied when I run lun show -v <volume>. Snaps = 200GB on that volume with 0% reserve. Here are the errors in the log file

vol is full (using or reserving 99% of space and 0% of inodes, using 99% of reserve).

Deleting snapshot 'hourly.0' in volume vol' to recover storage

Snap Delta hourly.1        hourly.0             114180092kb = 200gb

Even with Autodelete, it seemed to have went offline? But my eyes weren't on it when it was recovered so I may not have all the details.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
ok..
0
 
snyderkvAuthor Commented:
I meant 7TB volume

And you don't need a 7tb volume if you were going to play with it. It could be a 7gb volume and 5gb store. Oh I have vsim on my laptop so I'll see what I can find.
0
 
snyderkvAuthor Commented:
Deleting snapshot 'hourly.0' in volume 'vol' to recover storage
Insufficient space on volume 'vol" to perform operation. 64.0KB was requested but only 8.00KB was available.
LUN 'vol/rdm' has run out of space
LUN rdm has been taken offline

Here is the rest of the errors. You can see snapshot autodelete just deleted a snapshot before the rdm went offline. The RDM itself is 1TB on the windows filesystem and shows 5TB occupied on netApp "being an RDM" Autogrow was on but grew to it's max size. Audodelete settings are on default. I'll lab it tonight.
0
 
snyderkvAuthor Commented:
Quitting on this since my Ontap vsims crashed again. But what I did learn is that sis stores a copy of metadata in the volume and could have had stuff there as well as snaps. Also could have been fragmentation or free blocks. Maybe reallocation and reclamation would have helped. Anyways, moving on.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
ok...
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

  • 9
  • 6
Tackle projects and never again get stuck behind a technical roadblock.
Join Now