disk i/o latency on 1850 4e/si

[Andrew Reid]

Any chance you have an active snapshot of the partition under test on 
the LVM based system (on the same spindles). That would cause double the 
I/O and a lot more seeks.

How does the read performance compare.






J Potter wrote:
> Hi List,
> 
> Has anyone else encountered high latency for disk i/o on a PE 1850?  
> What was the solution?
> 
> We have a server that's been running fine for several years, and over  
> the past few months, i/o throughput has gotten dramatically worse.  
> I've been unable to track it down to any particular change.
> 
> We're running CentOS 4.7, relatively standard config. I've compared  
> bonnie++ output between this server and another 1850 running CentOS 4.  
> (See "vs" column below.) omreport shows both systems as having the  
> same write through settings on the raid controller. I've not been able  
> to find any clues in system logs or omreport logs. Both servers are a  
> raid-1 config, although the one with performance issues is running  
> LVM. Both servers use ext3 as the filesystem.
> 
> Sequential Output:
> 	Latency K/sec Per Char:  1379ms  vs  41ms  (33x)
> 	Latency K/sec Block:        3015ms  vs  846ms (3x)
> 
> Sequential Create:
> 	files /sec:		1926   vs   22816 (11x)
> 	Latency:			334ms  vs  0.2ms (1670x)
> 
> Unfortunately, the server is in production, so it's not so easy to  
> isolate things. The test results will be skewed because of other  
> processes running on the production instance.
> 
> Thoughts? I'm stumped.
> 
> -Jeff
> 
> _______________________________________________
> Linux-PowerEdge mailing list
> Linux-PowerEdge at dell.com
> http://lists.us.dell.com/mailman/listinfo/linux-poweredge
> Please read the FAQ at http://lists.us.dell.com/faq
>