Re: re; How do <>rw and <>rd_wr disk tests work?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2022/01/19 6:10, Gibson, Thomas wrote:
> Ok, so I've gotten info, from sitsofe@xxxxxxxxx, on the fio test data between
>       seqrd, seqwr, randrd, randwr.
> and
>       seqrw, seqrd_seqwr
>       randrw, randrd_randwr, randrd_seqwr.
> 
> That person that reviewed my test data, id'd that cpu usage was not an issue.  So, I'm now looking at my buffer cache size and what unneeded processes the server is running that I can turn off.  
> 
> I'm trying to determine what things on the server will affect the testing beyond what I'm currently addressing.  It was also suggested that I add "iostat -xzh 1" to see what disk utilization is like.
> 
> I know this is a biggy, does anyone know where I can find a copy of JEDEC JESD218?

These documents are for purchase or free to access for JDEC members. Either
purchase it or get access to it if HPE is a JEDEC member (likely it is).

> 
> But I'm still left with a basic question, are there any kind of documents that describe how the tests operate internally?  By understanding how the testing operates I might be able to identify what could explain the results I'm getting.

What test ? Can you share the fio script/command line you are running and what
your performance issues are ?
There are many ways to do IOs to storage. Without details on what exactly you
are running, it is impossible to comment on performance.

> 
> Tom Gibson
> hw/sw engineer
> hpe/aruba
> 
> 
> 
> 


-- 
Damien Le Moal
Western Digital Research



[Index of Archives]     [Linux Kernel]     [Linux SCSI]     [Linux IDE]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]

  Powered by Linux