Why LUN is using more space than allocated?

[fajarpri2]

Hi all,

vol1 300GB

Inside vol1 I have 1 LUN 300GB too.

I increase vol1 to 700GB

But then when I try to increase that LUN to 700GB, it says the maximum size is 500GB.

Doing df -h shows that vol1 is using more than 300GB

BEFORE:

n3600b> df -h
Filesystem               total       used      avail capacity  Mounted on
/vol/vol0/               537GB     1448MB      535GB       0%  /vol/vol0/
/vol/vol0/.snapshot      134GB      411MB      133GB       0%  /vol/vol0/.snapshot
/vol/vol1/               300GB      300GB        0KB     100%  /vol/vol1/
/vol/vol1/.snapshot        0KB     1284KB        0KB     ---%  /vol/vol1/.snapshot

AFTER:

n3600b> df -h
Filesystem               total       used      avail capacity  Mounted on
/vol/vol0/               537GB     1448MB      535GB       0%  /vol/vol0/
/vol/vol0/.snapshot      134GB      411MB      133GB       0%  /vol/vol0/.snapshot
/vol/vol1/               700GB      496GB      203GB      71%  /vol/vol1/
/vol/vol1/.snapshot        0KB     1284KB        0KB     ---%  /vol/vol1/.snapshot

Why is that? Who use the missing 200GB?

[fajarpri2]

Well, I suspect this is because of FR?

But why FR acts inconsistently?

On other volume, I can use 100% of the space allocated in the a volume to a LUN.

[tjlee · Accepted solution]

You're most likely correct in your assessment of "fractional reservation" being the cause.  To fully understand space reserve in a LUN environment, I'd recommend reading through TR-3483.

http://media.netapp.com/documents/tr-3483.pdf

Cheers, Tony

[fajarpri2]

Thanks Tony.

This command shows the missing 200GB.

n3600b> df -rh           
Filesystem               total       used      avail   reserved  Mounted on
/vol/vol0/               537GB     1449MB      535GB        0KB  /vol/vol0/
/vol/vol0/.snapshot      134GB      415MB      133GB        0KB  /vol/vol0/.snapshot
/vol/vol1/               700GB      700GB        0KB      198GB  /vol/vol1/
/vol/vol1/.snapshot        0KB     1412KB        0KB        0KB  /vol/vol1/.snapshot

Hmmmmm......

The PDF said that Netapp recommendation is to set FR 0%.

So I think it is SAFE because:

1. I don't use snapshot at all

2. I don't overcommit the LUN.

What do you think?

[tjlee]

Yes, you can set the FSR to 0% in that case.  You may also want to consider the following:

1) Doublecheck that the "snap sched" for that volume is indeed turned off "0"

2) Also as a safety net, you may want to consider turning on "snap autodelete" on the volume, just in case someone accidentally turns on snapshots or tries to take a snapshot

BTW, here's another great Communities post which provides details of the FSR in action: http://communities.netapp.com/blogs/ServiceBytes/2009/03/23/fsr-in-action

Cheers, Tony

[fajarpri2]

Ah, it turns out that although I have turned off snapshot on vol1, there were already some snapshots created.

So, right after I delete them, the reserve space is now 0 eventhough the FR is still 100%.

Thanks so much Tony for your kindness and the insight view.