Subscribe

volume size display error after deduplication

I have an interesting problem, I performed a deduplication on a volume that contained a space reserved LUN.

After doing much reading, this is effectively a configuration that while works, provides none of the benefits.

Further reading has people saying to turn off LUN space reservation when you have space guarantees set to volume at the FlexVol level.

This is all fine, however now that I have done that, the space the volume is reporting is inaccurate.

I have a volume that is 1TB in size and it is showing that it is almost completely free, even though I have a LUN on it, and at the host level that LUN contains considerable data:

na2*> df -h vmware_02

Filesystem               total       used      avail capacity  Mounted on

/vol/vmware_02/          999GB       53MB      999GB       0%  /vol/vmware_02/

/vol/vmware_02/.snapshot      249GB     2512KB      249GB       0%  /vol/vmware_02/.snapshot

na2*> df -sh vmware_02

Filesystem                used      saved       %saved

/vol/vmware_02/           53MB      174MB          76%

na2*> lun show /vol/vmware_02/v2

        /vol/vmware_02/v2            1.0t (1099529453568) (r/w, online, mapped)

As you can see at the VMware level, that LUN is 1TB that has 600gb used, yet clearly at the NetApp level it sees the volume as empty.

My first thoughts are that this is a bug, and that running deduplication on a volume that contained a space reserved LUN has tickled a bug.

Anyone have any insight here?

Re: volume size display error after deduplication

Please show output of “df -S” instead (or small “s”, I do not remember which version supports which flag).

volume size display error after deduplication

na2> df -S vmware_02

Filesystem                used       compressed         a-sis     %saved

/vol/vmware_02/          55192                0        178584        76%

Filesystem                used      saved       %saved

/vol/vmware_02/           53MB      174MB          76%

Re: volume size display error after deduplication

It would be interesting to see “lun show -v” to see actual space consumption (which version of Data ONTAP BTW?) and “aggr show_space”.

Are you absolutely sure it is really the same LUN and volume?

Re: volume size display error after deduplication

na2*> lun show -v /vol/vmware_02/v2

        /vol/vmware_02/v2            1.0t (1099529453568) (r/w, online, mapped)

                Serial#: Hn/3NZl6kAst

                Share: none

                Space Reservation: disabled

                Multiprotocol Type: vmware

                Maps: igroup_vmware_fc=8

Aggregate 'aggr6_sata'

    Total space    WAFL reserve    Snap reserve    Usable space       BSR NVLOG           A-SIS

   9546637056KB     954663680KB      85919732KB    8506053644KB             0KB        534644KB

Space allocated to volumes in the aggregate

Volume                          Allocated            Used       Guarantee

backups                       210880284KB      53531752KB          volume

vmware_nosnap_02             1187256000KB      55901480KB          volume

vmware_02                    1316947376KB         83956KB          volume

Aggregate                       Allocated            Used           Avail

Total space                  2715083660KB     109517188KB    5790117788KB

Snap reserve                   85919732KB      17870608KB      68049124KB

WAFL reserve                  954663680KB       7930032KB     946733648KB

na2*> version

NetApp Release 7.3.3: Thu Mar 11 22:29:52 PST 2010

Re: volume size display error after deduplication

Well … it looks like it the correct LUN (at least, serial number matches). Unfortunately, Data ONTAP is too old and does not show “Occupied size”.

Does VMware have something like “thin provisioning” too? Is it possible that it has 600GB reserved, but not actually used? IIRC it has different VMDK modes; one of them is “allocate but zero out on first access”, which sounds like it may be the case.

Have you tried to ask on VMware communities?