[linux-lvm] LVM iostat readings

[Lyn Rees]

Anyone?

__________________

Hi,

I've had a good look around for answers but found nothing useful.

I've collected data using iostat from a production server. I've got 3 
disks stitched together as a linear LVM, 250GB+500GB+500GB, with data on 
only the first 2 disks at the moment.

Under load, the PV's get busy showing high utilisation (%util) and quite 
high queue depths (avgqu-sz). the LV however hits 100% utilisation and the 
queue depth gets huge! Performance becomes poor at these times, suggesting 
the values are representative of what is actually happening.

Can someone help me understand what's going on here please? Could LVM be 
causing a bottleneck, or is a physical bottleneck causing this?

Thanks,
Lyn Rees.
--------------------------------------------------
Mr Lyn Rees
Senior Engineer, UIG
Information Services Computing Centre
Cardiff University, 40-41 Park Place,
Cardiff. CF10 3BB.
--------------------------------------------------
Contact numbers:
 (029) 2087 9188 (direct)
 (029) 2087 4875 (reception)
 (029) 2087 4285 (fax)
--------------------------------------------------
Email: rees at cardiff.ac.uk
Web: www.cardiff.ac.uk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-lvm/attachments/20090305/5b65b170/attachment.htm>