using or reserving 98% of space and 0% of inodes, using 63% of reserve

[TDUBB1234]

So I still trying to understand how netapp disk space is allocated ....

I created a volume in an aggregate

vol create myvol myaggr 100g

---------------

then create a lun for 50gb

lun create -s 50g -t vmware /vol/myvol/mylun

                              total         used        available

/vol/myvol/               80GB       50GB       29GB      63%  /vol/myvol/

/vol/myvol/.snapshot       20GB        0GB       19GB       0%  /vol/myvol/.snapshot

---------------------

Then mounted the lun and moved a vm ~30gb vm to this lun

afterwards

I got this message

: monitor.globalStatus.nonCritical:warning]: /vol/myvol is full (using or reserving 98% of space and 0% of inodes, using 63% of reserve).

df -V -g shows

/vol/myvol/               80GB       80GB        0GB     100%  /vol/myvol/

/vol/myvol/.snapshot       20GB        0GB       19GB       0%  /vol/myvol/.snapshot

---------------------

and in esx the actual size of the lun is using

vmfs3                    49.8G     32.7G     17.1G  66% /vmfs/volumes/4e7a733c-8cf0ad8a-b1fe-001b2122b15d

------------------------

so I started with a 50gb lun in a 80gb volume. then i wrote 32.7gb into the 50gb lun.

what i still dont understand is why

df -V -g shows

/vol/myvol is 80gb full now and 0gb available now

why not 17.1 gb available instead?

[aborzenkov]

Do you have any snapshots on this volume? "snap list myvol"?

Do you happen to know how VMware initializes datastore? I.e. - would it touch every block in a LUN during format?

[TDUBB1234]

snap list myvol

Volume myvol

working...

  %/used       %/total  date          name

----------  ----------  ------------  --------

  1% ( 1%)    0% ( 0%)  Sep 22 16:00  hourly.0      

  1% ( 1%)    0% ( 0%)  Sep 22 12:00  hourly.1      

  2% ( 0%)    0% ( 0%)  Sep 22 08:00  hourly.2      

  2% ( 1%)    1% ( 0%)  Sep 22 00:00  nightly.0     

  3% ( 1%)    1% ( 0%)  Sep 21 20:00  hourly.3      

  3% ( 0%)    1% ( 0%)  Sep 21 16:11  hourly.4      

[aborzenkov]

And what does "df -rh myvol" show?

[TDUBB1234]

df -rh myvol

Filesystem total used avail reserved Mounted on

/vol/myvol/ 80GB 80GB 0GB 30GB /vol/myvol/

/vol/myvol/.snapshot 20GB 963MB 19GB 0MB /vol/myvol/.snapshot

why is used 80GB?

lun is 50GB

why is reserved 30GB?

[aborzenkov]

By default when snapshot is created on volume with LUN Data ONTAP reserves amount of space that is “frozen” in snapshot on volume to ensure LUN still has space to write into it. Please read TR-3483 which explains all this in great details.

In general

- Snap reserve is meaningless on volumes containing LUN and simply reduce useable space

- Automatic (filer initiated) snapshots of volumes with LUNs will often result in inconsistent data on LUNs; the effect varies from hard to recover up to impossible to recover data from snapshot. Usually snapshots are initiated from host in coordination with applications on LUNs. You have automatic snapshots turned on.

[TDUBB1234]

i copied a vm to this volume. then removed it

from the esx host, the lun is showing as empty

but on the filer

df -V -g myvol

Filesystem               total       used      avail capacity  Mounted on

/vol/myvol/            80GB       61GB       18GB      76%  /vol/myvol/

/vol/myvol/.snapshot       20GB       30GB        0GB     154%  /vol/myvol/.snapshot

lun show -v /vol/myvol/foo

        /vol/myvol/foo             50g (53687091200)   (r/w, online, mapped)

                Serial#: dfkpVZfRtWuT

                Share: none

                Space Reservation: enabled (not honored by containing Aggregate)

                Multiprotocol Type: vmware

                Maps: test=200

                Occupied Size:   30.3g (32552378368) 

                Creation Time: Sun Sep 25 18:50:24 PDT 2011

                Cluster Shared Volume Information: 0x0

why is showing 61gb used and occupied size 30.3gb?