Skip to contentSkip to conversation

diff. in striping between HP-UX and Linux LVM

linux-lvm Stephan Hendl 2 participants 2 messages Sep 14, 2001

lore
Share & tools
lore
#0 lore raw

Hi all,

I have to work with both systems an found some differences in the striping between the two kinds of LVM. I created a vg consisting of two disks on the HP-UX and on the linux system and a lv with the "-i 2 -I 64"-option.

In the case of HP-UX the first LE points to the first PE on the first disk, the second LE points to the first PE of the second disk and so on.

In case of Linux LVM the LEs corrospond to the PEs, there is no alternation and therefore no enhancement of performance because only one disk will be used at the same time.

Where I am wrong?

ciao
Stephan

HP-UX:
--- Logical volumes ---

LV Name                     /dev/vgmail/lvol1
VG Name                     /dev/vgmail
LV Permission               read/write
LV Status                   available/syncd
Mirror copies               0
Consistency Recovery        MWC
Schedule                    striped
LV Size (Mbytes)            4096
Current LE                  1024
Allocated PE                1024
Stripes                     2
Stripe Size (Kbytes)        64
Bad block                   NONE
Allocation                  strict
IO Timeout (Seconds)        default

--- Distribution of logical volume ---

   PV Name            LE on PV  PE on PV
   /dev/dsk/c4t6d1    512       512
   /dev/dsk/c6t6d2    512       512

--- Logical extents ---

   LE    PV1                PE1   Status 1
   00000 /dev/dsk/c4t6d1    00000 current
   00001 /dev/dsk/c6t6d2    00000 current
   00002 /dev/dsk/c4t6d1    00001 current
   00003 /dev/dsk/c6t6d2    00001 current
   00004 /dev/dsk/c4t6d1    00002 current
   00005 /dev/dsk/c6t6d2    00002 current
   00006 /dev/dsk/c4t6d1    00003 current
   00007 /dev/dsk/c6t6d2    00003 current

...

Linux:
--- Logical volume ---

LV Name                /dev/vg00/lvol1
VG Name                vg00
LV Write Access        read/write
LV Status              available
LV #                   1
# open                 1
LV Size                4 GB
Current LE             1024
Allocated LE           1024
Stripes                2
Stripe size (KByte)    64
Allocation             next free
Read ahead sectors     120
Block device           58:0

--- Distribution of logical volume on 2 physical volumes ---

   PV Name                  PE on PV     reads      writes
   /dev/sdb1                512          1536487    1472056
   /dev/sdc1                512          1535960    1472315

--- logical volume i/o statistic ---
3072447 reads 2944371 writes

--- Logical extents ---

   LE    PV                        PE     reads      writes
   00000 /dev/sdb1                 00000  0          1008
   00001 /dev/sdb1                 00001  0          1024
   00002 /dev/sdb1                 00002  0          590
   00003 /dev/sdb1                 00003  0          848
   00004 /dev/sdb1                 00004  0          16
   00005 /dev/sdb1                 00005  0          0
   00006 /dev/sdb1                 00006  272        2594

--
LDS Brandenburg
Dr. Stephan Hendl
fon: +49-(0)331-39 471
fax: +49-(0)331-27548 1187
EMail: [email protected]

#1 lore raw

Stephan,

the alternation of the subalocated stripes isn't visible in "lvdisplay -v".

In principle the allocation looks like (assuming an exmaple LV with 3 stripes and 6 LE):

LE0 -> PEa on PV1
LE1 -> PEb on PV1

LE2 -> PEc on PV2
LE3 -> PEd on PV2

LE4 -> PEe on PV3
LE5 -> PEf on PV3

If the stripes have a size of 64K the accesses to the address space of the LV go like:

1st 64k -> 1st 64K PEa
2nd 64k -> 1st 64K PEc
3rd 64k -> 1st 64K PEe
4th 64k -> 2nd 64K PEa
5th 64k -> 2nd 64K PEc
6th 64k -> 2nd 64K PEe
...

Regards,
Heinz -- The LVM Guy --

Quoted 94 lines · Stephan Hendl

Hi all,

I have to work with both systems an found some differences in the striping between the two kinds of LVM. I created a vg consisting of two disks on the HP-UX and on the linux system and a lv with the "-i 2 -I 64"-option.

In the case of HP-UX the first LE points to the first PE on the first disk, the second LE points to the first PE of the second disk and so on.

In case of Linux LVM the LEs corrospond to the PEs, there is no alternation and therefore no enhancement of performance because only one disk will be used at the same time.

Where I am wrong?

ciao
Stephan

HP-UX:
--- Logical volumes ---

LV Name                     /dev/vgmail/lvol1
VG Name                     /dev/vgmail
LV Permission               read/write
LV Status                   available/syncd
Mirror copies               0
Consistency Recovery        MWC
Schedule                    striped
LV Size (Mbytes)            4096
Current LE                  1024
Allocated PE                1024
Stripes                     2
Stripe Size (Kbytes)        64
Bad block                   NONE
Allocation                  strict
IO Timeout (Seconds)        default
   --- Distribution of logical volume ---
   PV Name            LE on PV  PE on PV
   /dev/dsk/c4t6d1    512       512
   /dev/dsk/c6t6d2    512       512

   --- Logical extents ---
   LE    PV1                PE1   Status 1
   00000 /dev/dsk/c4t6d1    00000 current
   00001 /dev/dsk/c6t6d2    00000 current
   00002 /dev/dsk/c4t6d1    00001 current
   00003 /dev/dsk/c6t6d2    00001 current
   00004 /dev/dsk/c4t6d1    00002 current
   00005 /dev/dsk/c6t6d2    00002 current
   00006 /dev/dsk/c4t6d1    00003 current
   00007 /dev/dsk/c6t6d2    00003 current

...

Linux:
--- Logical volume ---

LV Name                /dev/vg00/lvol1
VG Name                vg00
LV Write Access        read/write
LV Status              available
LV #                   1
# open                 1
LV Size                4 GB
Current LE             1024
Allocated LE           1024
Stripes                2
Stripe size (KByte)    64
Allocation             next free
Read ahead sectors     120
Block device           58:0
   --- Distribution of logical volume on 2 physical volumes  ---
   PV Name                  PE on PV     reads      writes
   /dev/sdb1                512          1536487    1472056
   /dev/sdc1                512          1535960    1472315

   --- logical volume i/o statistic ---
   3072447 reads  2944371 writes

   --- Logical extents ---
   LE    PV                        PE     reads      writes
   00000 /dev/sdb1                 00000  0          1008
   00001 /dev/sdb1                 00001  0          1024
   00002 /dev/sdb1                 00002  0          590
   00003 /dev/sdb1                 00003  0          848
   00004 /dev/sdb1                 00004  0          16
   00005 /dev/sdb1                 00005  0          0
   00006 /dev/sdb1                 00006  272        2594

--
LDS Brandenburg
Dr. Stephan Hendl
fon: +49-(0)331-39 471
fax: +49-(0)331-27548 1187
EMail: [email protected]

Quoted 5 lines

_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

*** Software bugs are stupid.
Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen Sistina Software Inc.
Senior Consultant/Developer Am Sonnenhang 11
56242 Marienrachdorf Germany
[email protected] +49 2626 141200

                                                       FAX 924446

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-