dealing with PE and max pe per pv

[deepan_1]

hi gurus

just now i have started to learn lvm and i heard abt this forum site which makes new learner to easily solve their queries.

my doubt is

currently there are 2 pv
max pe per pv 3071
max pe 16
max pv 16
lvm version : v1

when i compared the previous post i had a doubt

max pe per pv * pe size === 3071*16 =49136==47 gb

max pe per pv * pe size * max pv ==786176==767 gb.

so that i can add 767 gb disk to that vg.

1.but my doubt is whether 767gb disk will fulli utilized or not since
2.max pe per pv is for single pv sicnce the term max pe per pv, but max per pv is not related to any pv since if we are adding any pv the size wont get increased unless we use vgmodify,
3.bdf output

/dev/vgsystem/lvoldatasystem
83853312 69363027 13584830 84 % /system/systemtest

nearly 66% of disk is used when i compared from bdf output, but max pe per pv * pe size === 3071*16 =49136==47 gb

how can a 47 gb disk space that can only be utilized is being used by 66 gb disk

thanks and regards

deepan

[P Arumugavel]

hi deepan,

No concerns about PV size in your case.
pe_size is 16MB
max_pe is 3071

Volume group version 1.0 helps you. max_pe * pe_size is PV size would be utilised. If the actual size of PV exceeds more than value of max_pe * pe_size, then max_pe is adjusted to match the value of physical volume size. But maximum number of physical extents can be a value in the range 1 to 65535.

Now lets come to your case:
>> max pe per pv * pe size === 3071*16 =49136==47 gb
>> max pe per pv * pe size * max pv ==786176==767 gb.

If your PV size exceeds the size of 47 gb, max_pe (3071) will get adjusted to match the original value of 767GB. But max_pe value would be limited with 1 - 65535 range.

So your max_pe will be max_pe*16MB=767GB as per your usage.
So you can add the disk size of upto 1048.56GB, which can be addressed by VG 1.0..

Hope, it clears you...

[TwoProc]

I disagree - unless that's a new OS feature in 11.3 or 11.2, I don't ever remember that the max pe per pv being automatically adjusted, in fact I remember quite the opposite.

If it's a new OS feature, then that's awesome, and I'm quite happy to hear about it.

[chris huys_4]

Hi Deepan,

You could just cut&paste what was in the attachment directly in the call, as the difference with what is displayed, is not much.

root / >vgdisplay -v vgsystem

--- Volume groups ---

VG Name /dev/vgsystem

VG Write Access read/write

VG Status available, exclusive

Max LV 255

Cur LV 1

Open LV 1

Max PV 16

Cur PV 2

Act PV 2

Max PE per PV 3071

VGDA 4

PE Size (Mbytes) 16

Total PE 5118

Alloc PE 5118

Free PE 0

Total PVG 0

Total Spare PVs 0

Total Spare PVs in use 0

VG Version 1.0

VG Max Size 786176m

VG Max Extents 49136



--- Logical volumes ---

LV Name /dev/vgsystem/lvoldatasystem
LV Status available/syncd

LV Size (Mbytes) 81888

Current LE 5118

Allocated PE 5118

Used PV 2





--- Physical volumes ---

PV Name /dev/disk/disk120

PV Status available

Total PE 3071

Free PE 0

Autoswitch On

Proactive Polling On



PV Name /dev/disk/disk541

PV Status available

Total PE 2047

Free PE 0

Autoswitch On

Proactive Polling On


/dev/vgsystem/lvoldatasystem

83853312 69363027 13584830 84% /system/systemtest


> how can a 47 gb disk space that can only be
> utilized is being used by 66 gb disk
The volumegroup has 2 pvs, 1 pv of 47Gbyte and 1 pv of 2047 * 16m =34 GB and the logical volume "behind the system/systemtest mountpoint" is using the 2 pv's. So its normal that the filesystem has 81GB available..

Greetz,
Chris