Hi ,
before I opened a new thread here I did some research on the documentation and internet - but I couldn't find an explanation.
when running for example a "sar -dp 2 22" , I see that there are way more tps on a particular lvm volume than on the underlying disk.
There is one disk /dev/sdd which is used as physical volume for a volume group PPWP ( see below ). Several logical volumes where created of that volume group.
The Filesystem on the logical volumnes is ext3. When the database is doing IO on the lv , sar shows more tps on the lv than on the pv ( = /dev/sdd )
Code:
[serv1:/dev/mapper]sar -d 2 2
Linux 2.6.18-164.el5 (serv1) 04/05/2014
07:57:29 AM DEV tps rd_sec/s wr_sec/s avgrq-sz avgqu-sz await svctm %util
07:57:31 AM dev8-0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev8-1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev8-2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev8-16 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev8-32 1.00 0.00 32.00 32.00 0.00 0.00 0.00 0.00
07:57:31 AM dev8-48 881.50 582584.00 3743.00 665.15 6.93 7.97 1.14 100.50 <<<< /dev/sdd Disk
07:57:31 AM dev253-0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev253-1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev253-2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev253-3 6.00 0.00 1833.50 305.58 0.01 1.67 1.67 1.00
07:57:31 AM dev253-4 6.50 0.00 1813.50 279.00 0.01 1.54 1.54 1.00
07:57:31 AM dev253-5 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
07:57:31 AM dev253-6 2257.00 580016.00 96.00 257.03 15.57 6.94 0.45 100.50 <<< Logical volume PPWP-lvdata
07:57:31 AM dev253-7 1.00 0.00 32.00 32.00 0.00 0.00 0.00 0.00
As you can see dev253-6 has more than twice tps than the underlying disk
This is the disk (dev8-48)
[serv1:/dev/mapper]ll /dev/sdd
brw-r----- 1 root disk 8, 48 Feb 11 14:03 /dev/sdd
This is the logical volme (dev253-6)
[serv1:/dev/mapper]ll /dev/mapper/
total 0
crw------- 1 root root 10, 63 Nov 20 19:56 control
brw------- 1 root root 253, 7 Nov 20 19:57 PPWPFRA-lvfra
brw------- 1 root root 253, 6 Nov 20 19:57 PPWP-lvdata <<<<<< dev253-6
brw------- 1 root root 253, 5 Nov 20 19:57 PPWP-lvoracle
brw------- 1 root root 253, 3 Nov 20 19:57 PPWP-lvredo1
brw------- 1 root root 253, 4 Nov 20 19:57 PPWP-lvredo2
brw------- 1 root root 253, 0 Nov 20 19:57 VolGroup00-root
brw------- 1 root root 253, 2 Nov 20 19:56 VolGroup00-swap
brw------- 1 root root 253, 1 Nov 20 19:57 VolGroup00-var
This is the physical volume ( disk /dev/sdd ) - the only disk in this volumegroup
[serv1:/dev/mapper]pvs
PV VG Fmt Attr PSize PFree
/dev/sda2 VolGroup00 lvm2 a- 89.88G 0
/dev/sdb lvm2 -- 54.00G 54.00G
/dev/sdc PPWPFRA lvm2 a- 150.00G 96.00M
/dev/sdd PPWP lvm2 a- 950.01G 16.00M <<<<<
My first assumption was that the filesystem cache could be the reason, but the database running on this system is doing DIRECT IO. So I ruled that out.
Thanks for any explanations or guesses !
best regards