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.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
server HDD bottleneck?
Hi everyone,
from some time I have taken out a server that I had in the basement with a pair of Xeon e5 2620 v1, I bought four barracuda hard disks and configured them in raid 10.
The use that I have to do is to let us run about ten windows vps, but i'm actually getting the VPS really slow, windows installations takes 2-3 hours and the applications inside windows keeps crashing.
I was wondering if is this a HDD bottleneck, if someone have faced similar problems, please let me know your experience.
Comments
Wich RAID Controller is used in your server?
I also use Barracuda HDDs in some servers and didn't got any performance problems.
Welcome to LowEndSupport.
Do you also need 24/7 Dedicated support for another $696,9/m ?
What kind of Raid card? With BBU? Or software?
Please atleast mention your host OS, virtulizor you are using and if proxmox, have you got ballooning device and VirtIO configured on gues VMs ?
Centos 6,
SolusVM
Raid card: HP P420 (1GB FBWC)
How much RAM does the server have? And how much RAM do you give to each Windows VM?
128 GB ram, and I used like 50 GB at the moment
someone know what it can be?
Is the ram any good? memtest it.
Edit: Just install Linux and put a lot of OpenVZ NAT VPSes with no swap whatsoever.
I did some userbenchmark on the Windows VPS and the RAM was good. (about 11 GB/s)
The apps crashing confused me.
Assuming everything is great from software point of view, random app crashes could be due to memory corruption.
It's not about the speed, it's about memory chips gone bad due to usage, cosmic rays or whatever...
But when i think about it, it's not that likely and quite likely had that been the case, you would have already gotten ECC error warnings somewhere.
The apps crashed after a long loading time, for example yesterday was trying to install windows and I got "Windows stopped working" during the boot process.
What does S.M.A.R.T. says? Are the HDDs done for?
Most likely drivers if not hardware. Upgrade firmware, perform cpu, ram and hdd stress tests.
Maybe a virtualization issue. You should use virtio drivers for disks and i reach nearly native IO using native io mode.
https://docs.solusvm.com/display/DOCS/1.20
To rule out a virtualization problem, you should simply do some IO tests directly on the host system.
First benchmark the drive array from the node itself (CentOS). Also: what does the Raid controller firmware say? E.g., is it fully done with initialization and reports the array as OK and operational?
I agree with @Janevski, disks could be damaged, could you run smartctl command and check disk age/errors ?
SMART Health Status: OK
Trying to do the drive test:
Long (extended) offline self test failed [unsupported scsi opcode]
Windows had some drive issues... Check BIOS ide/sata/raid compatibility mode on, something something...
Also check the installation media, the dvd might be corrupted, dvd drive might be dusty as well, or usb flash might be not properly prepared, or even the flash itself might have memory issues...
Off topic, inspect the board for any capacitor damage.
The windows has been loaded via KVM, I didn't used any dvd drive etc..
Hi everyone, thanks for the past help, I noticed that the SusWrite is really slow on userbenchmark.
I'm getting around 150 MB/s Read/Write
but only 9 MB/s in SusWrite
This is the only apparent problem on the server, someone know to what it can be related?
Thanks.
Just curious, do you have the RAID battery or supercapacitor installed? If yes, check the battery/supercapacitor status?