Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Server has SSD but showing ROTA 1
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Server has SSD but showing ROTA 1

ViridWebViridWeb Member, Host Rep

Hey guys!
Recently we have bought few servers to use in our rack in serverius DC

on of the server has 2 x Intel Xeon Gold CPU
4 x 3.84TB Intel S4510 SSD
It also has Hard RAID 10 using LSI 9361-4i and BBU

But the read and write speed of the server is very poor most of the time.
Even we have create few VPS on the server but VPS also showing poor read and write speed.

So we have check the following command to get ROTA result and it's showing ROTA 1

[root@vm01 ~]# cat /sys/block/sda/queue/rotational
1
[root@vm01 ~]# lsblk -d -o name,rota
NAME ROTA
sda 1

It should return 0 if it's SSD instead 1 right?

Today the DC confirm it has 4 x 3.84TB Intel S4510 SSD

They said maybe there some issue with RAID config, or OS install or Bios

We are checking but still not getting any result/issue till now.

So my question is there anyone who face same kind of situation?

Any suggestion?

Thank You!

Comments

  • WebProjectWebProject Host Rep, Veteran

    [root@vm01 ~]# cat /sys/block/sda/queue/rotational
    1
    [root@vm01 ~]# lsblk -d -o name,rota
    NAME ROTA
    sda 1

    Should be zero (0):

    info from sysfs

    cat /sys/block/sde/queue/rotational

    1
    You should get 1 for hard disks and 0 for a SSD.

    lsblk

    lsblk -d -o name,hctl,rota

    NAME HCTL ROTA
    sda 0:0:0:0 1
    sdb 0:0:1:0 1
    sdc 0:0:2:0 1
    1 for hard disks and 0 for a SSD.

    The datacentre probably is looking at different server, check the following command:
    smartctl -a /dev/sda | grep Rotation

    as it will tell you the hard drive is solid state device or hdd rpm speed.

  • ViridWebViridWeb Member, Host Rep

    Ok the issue of ROTA solve as we attached these disks with RAID controller that's why it's showing ROTA 1

    But still don't know why the speed is poor.
    Using RAID 10 so it must be increase more

    official speed of these drives are

    Sequential Read (up to) 560 MB/s
    Sequential Write (up to) 510 MB/s
    Random Read (100% Span) 97000 IOPS
    Random Write (100% Span) 32000 IOPS
    Latency - Read 36 µs
    Latency - Write 37 µs
    Power - Active 3.6W
    Power - Idle 1.1W

    And I am getting like this
    Write Speed

    sync; dd if=/dev/zero of=tempfile bs=1M count=1024; sync
    1024+0 records in
    1024+0 records out
    1073741824 bytes (1.1 GB) copied, 1.67986 s, 639 MB/s

    Read Speed (Maximum)

    dd if=tempfile of=/dev/null bs=1M count=1024
    1024+0 records in
    1024+0 records out
    1073741824 bytes (1.1 GB) copied, 0.202361 s, 5.3 GB/s

  • ViridWebViridWeb Member, Host Rep

    @WebProject said:

    [root@vm01 ~]# cat /sys/block/sda/queue/rotational
    1
    [root@vm01 ~]# lsblk -d -o name,rota
    NAME ROTA
    sda 1

    Should be zero (0):

    info from sysfs

    cat /sys/block/sde/queue/rotational

    1
    You should get 1 for hard disks and 0 for a SSD.

    lsblk

    lsblk -d -o name,hctl,rota

    NAME HCTL ROTA
    sda 0:0:0:0 1
    sdb 0:0:1:0 1
    sdc 0:0:2:0 1
    1 for hard disks and 0 for a SSD.

    The datacentre probably is looking at different server, check the following command:
    smartctl -a /dev/sda | grep Rotation

    as it will tell you the hard drive is solid state device or hdd rpm speed.

    It's true but DC said they have double checked and we have 4 SSD on it and everything matched what we have send to them

    but our Admin said

    It is true that SDA block rotational value should be 0 if it is a SSD drive; however, on this server the disks are not directly attached to the server. The disks are attached to the RAID controller and then this RAID controller projects make the array projects one to the operating system, so the SDA block rotation value will never be true in this case.

  • ViridWebViridWeb Member, Host Rep

    lsblk -d -o name,hctl,rota
    NAME HCTL ROTA
    sda 0:2:0:0 1

    It's also showing ROTA 1

    But DC confirm twice that we have 4 SSD present

  • ClouviderClouvider Member, Patron Provider

    sda is not the SSD. It’s your assembled raid array spit out to the system. You’re paying too much attention to whether something rotated or not, that’s unhealthy;-).

    Issue is your raid controller. Adjust stuff there. Might need to change arrray config. Controller config. Or the controller all together if too slow cache:

    Thanked by 2jsg seriesn
  • ViridWebViridWeb Member, Host Rep

    @Clouvider said:
    sda is not the SSD. It’s your assembled raid array spit out to the system. You’re paying too much attention to whether something rotated or not, that’s unhealthy;-).

    Issue is your raid controller. Adjust stuff there. Might need to change arrray config. Controller config. Or the controller all together if too slow cache:

    What should be the expected speed?
    I mean 639Mb/s write speed is too low for SSD RAID 10

  • ClouviderClouvider Member, Patron Provider

    That depends. You’ll need to re-search best practises on this controller before you find out.

  • deankdeank Member, Troll
    edited September 2019

    I've seen everything. @Cyder just gave someone a mental health advice.

  • jsgjsg Member, Resident Benchmarker

    @ViridWeb

    Listen to @Clouvider and have a closer look at your Raid controller as well as the driver.

  • ViridWebViridWeb Member, Host Rep

    @jsg said:
    @ViridWeb

    Listen to @Clouvider and have a closer look at your Raid controller as well as the driver.

    Yes we are still checking the controller even the the BIOS as well..
    Hope we can find the issue soon..

  • jsgjsg Member, Resident Benchmarker

    @ViridWeb said:

    @jsg said:
    @ViridWeb

    Listen to @Clouvider and have a closer look at your Raid controller as well as the driver.

    Yes we are still checking the controller even the the BIOS as well..
    Hope we can find the issue soon..

    I wouldn't worry about the (system) BIOS. Look at the Raid controller and also at the drivers!

  • AlexBarakovAlexBarakov Patron Provider, Veteran

    I'd start by updating the firmware of whatever you can. Proceed with making the RAID card as HBA/pass through mode and checking each drive individually.

    Thanked by 1jsg
  • ViridWebViridWeb Member, Host Rep

    @AlexBarakov said:
    I'd start by updating the firmware of whatever you can. Proceed with making the RAID card as HBA/pass through mode and checking each drive individually.

    Thanks for the suggestion but I have checked and firmware already up to date

    But still it's not clear why I'm getting only 639Mb/s write speed on main server and only 179Mb/s on a VPS on it

    Because I thik the speed should be more higher than that this because it's on SSD RAID 10
    Even for both main server and vps on it

  • ClouviderClouvider Member, Patron Provider

    You're not going to get your issue resolved for yourself on a forum. You should really invest into some competent assistance, especially as it appears you're charging money for your services.

  • jsgjsg Member, Resident Benchmarker

    @ViridWeb said:
    But still it's not clear why I'm getting only 639Mb/s write speed on main server and only 179Mb/s on a VPS on it

    Because I thik the speed should be more higher than that this because it's on SSD RAID 10
    Even for both main server and vps on it

    and

    @Clouvider said:
    You're not going to get your issue resolved for yourself on a forum. You should really invest into some competent assistance, especially as it appears you're charging money for your services.

    While I'd see and put it a bit less harsh @Clouvider is right. Main reason (and please don't take this personal. It's absolutely not): You are obviously clueless even to the point of not knowing what's relevant and how to measure.

    For a start those are almost certainly two unrelated problems. One is about the controller, the other one is highly likely about how you configured your hypervisor/VPSs.
    Focus on the controller first. Once it's OK you can concentrate on the hypervisor.

    As for the controller your measurements (although not exactly sensible) suggest that your Raid controller is working more or less properly. Unfortunately the information you provide does not allow us to understand the details better. Please note that there are many and inter-related factors and parameters to look at and to properly configure.
    And once more: also have a good look at the drivers. Usually there are quite some screws to turn and that can make a big difference.

  • XiNiXXiNiX Member, Host Rep
    edited September 2019

    May be try this :

    # sginfo -i /dev/sde | grep -i product
    Product : HITACHI HUA7210S
    
Sign In or Register to comment.