[CentOS] slow performance on company production server I need help

Wed Apr 22 16:00:55 UTC 2020
Gordon Messmer <gordon.messmer at gmail.com>

On 4/22/20 8:53 AM, Christopher Wensink wrote:
> I had an 8 TB External USB disk plugged into the system, that I had been
> using for additional space for backups, I was under the impression that
> sda, sdb, sdc, and sdd were the four disks on the raid controller card,


Not exactly.  If you have a RAID5 array, then you have one volume spread 
across the physical disks.  You can then divide that into smaller 
virtual disks, each of which is still spread across disks.  The OS 
doesn't see the four component disks, directly.


> but after unplugging the usb drive when running hdparm I am getting this:
>
> [root at daisy dev]# hdparm -tT /dev/sdb
>
> /dev/sdb:
> read() hit EOF - device too small


sdb was the disk that appeared in the kernel errors, so I'd imagine that 
you've fixed the problem by removing the USB drive enclosure.


> I'm still trying to find the exact smartctl command syntax to make it work.


https://www.cyberciti.biz/faq/unix-linux-freebsd-3w-9xxx-smartctl-check-hard-disk-command/