"LogicalExtentsNUmber" is bigger than the maximum value allowed

[Penni Nussbaum]

I have an Logical volume that is not striped. I tried to increase the size by 20 gig nd recived the message in the title. Why?

--- Logical volumes ---
LV Name /dev/vg03/lvol5
VG Name /dev/vg03
LV Permission read/write
LV Status available/syncd
Mirror copies 0
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 30152
Current LE 7538
Allocated PE 7538
Stripes 0
Stripe Size (Kbytes) 0
Bad block on
Allocation strict
IO Timeout (Seconds) default

[James R. Ferguson]

Hi Penni:

You have exceeded the number of physical extents allowed for the volume group. If you do a 'vgdisplay' for the volume group to which the logical volume belongs, and look at the value for "Total PE", you will see the maximum *number* of extents you can allocate on the volume group. Multiply this by the "PE Size" [by default, 4MB] to obtain the size (in MB) of the largest logical volume you can allocate.

...JRF...

[Penni Nussbaum]

I am confused. The vgdisplay shows the Total PE as 25884 . The PE Size (Mbytes) is 4. This makes 103536. However the bdf shows a bigger size.
/dev/vg03/lvol5 30875648 12987063 16770567

[A. Clay Stephenson]

Hi Penni:
bdf is reporting in 1K blocks; you need to multiply your 103536 X 1024 to get to 1K blocks. There is no discrepancy.

Clay

[James R. Ferguson]

Hi Penni:

The 'bdf' units are KB.

1GB = 1,000 MB = 1,000,000 KB.

You show a ~30.88 GB filesystem.

When you did your 'lvextend' if you used '-L ' then the value is *megabytes*. If you use a lowercase "l" as '-l ' then the is *extents*. It is a common mistake to mixup the letter argument leading to trying to specify a giant number of extents when one means megabytes. The largest is 65,535 (extents).

...JRF...

[Penni Nussbaum]

O.K., I'm a little thick

Total PE for this volume group: 25884
PE size 4
total 103536
times 1024
total 106020864

This still doesn't correspond to my bdf

/dev/vg03/lvol5 30875648 12979006 16778121

[Christopher McCray_1]

I may be in left field here, but what is your available PE/LE on the vgdisplay? The number of PE/LE on the vgdisplay is the total for all lvols on the vg and don't correspond to your lvdisplay/bdf, It looks like you have 30GB allocated to /dev/vg03/lvol5 on total of 103 GB for vg03. Does this clear anything up? You have used all you can allocated to the lvol.

Hope this helps.

Chris

[Marcin Wicinski]

Hi Penni,

give us Max PE, and the capacity of your disk.

[Penni Nussbaum]

Vgidpsplay is below. These are emc disks configured as 9 gig hypervolumes.

VG Name /dev/vg03
VG Write Access read/write
VG Status available
Max LV 255
Cur LV 13
Open LV 13
Max PV 16
Cur PV 12
Act PV 12
Max PE per PV 2157
VGDA 24
PE Size (Mbytes) 4
Total PE 25884
Alloc PE 19413
Free PE 6471
Total PVG 0

[Christopher McCray_1]

Just following up; did you understand the difference that myself, James and Clay are trying to convey:

1) You executed an lvextend to increase your filesystem by 20 GB and the output of your lvdisplay shows your LV size as 30152Mb current LE/allocated PE = 7538(Mb) (30 GB), which corresponds to your bdf output of 30875648 (7538 x 4(Mb) = 30152 x 1024 = 30875648 1Kb blocks)

2) Your vgdisplay shows that your max PE is 25884 ( 25584(Mb) x 4(Mb size) = 103536(Mb))
This is the size(Mb) of your disk(s) within you vg03. The two sizes are mutually exclusive of each other.

Going back to the beginning, I really am not sure what the big question is, but I hope I'm on the right track. Good luck

Chris

[Alan Riggs]

Please give the exact text of the command that is failing. I suspect that a previous response might be correct, you are confusing -l with -L.

On another note--you say these volumes are onj EMC drives, but your lvdicplay shows bad block relocation ON. This is a mistake. Logical volumes created on EMC hypers (or most other "intelligent" disk arrays) should not use system level bad block relocation. Unfortunately, I know of no way to correct that short of recreating the logical volumes and repopulating data.

[James R. Ferguson]

Hi Alan:

You can use 'lvchange' to turn off the bad block reallocation on EMC2 disk. I've done this after-the-fact.

Regards!

...JRF...

[Alan Riggs]

Good news, James. I have never had to do this personally, but it's nice to know the capability is there.

Penni, I strongly suggest that you turn bad block relocation to N for all logical volumes residing on EMC hypers.