I/O Errors: Supermicro 1027R-N3RF, Samsung SSD 840 Pro MZ-7PD512BW, Areca ARC-1680

[sergio.di]

Hi

I've got some big problems with my proxmox setup.

Generally all works fine now, at the beginning i got some more problems with software-raid setup. Now I've changed that to an Areca ARC-1680 hardware RAID-controller with 4 Intel SSDs in a RAID 5 setup, which seems to work fine...for the first.. Everything is running on a Supermicro 1027R-N3RF Server.

Now there is a problem when copying some big files/many files at same time, i.e. when cloning a machine.

In /var/log/messages i get errors like the following

[INDENT]kernel: arcmsr0: abort device command of scsi id = 0 lun = 0 
kernel: arcmsr0: wait 'abort all outstanding command' timeout 
kernel: arcmsr0: executing hw bus reset .....
kernel: Areca RAID Controller0: F/W V1.51 2012-07-04 & Model ARC-1680
kernel: arcmsr0: abort device command of scsi id = 0 lun = 0
[/INDENT]

Proxmox then is completely unuasable. I/O-Load is too high to do something other..every other thingts u wanna do in that moment doesnt work, not even a ssh-connection i.e.

Does anyone has some experience with this kind of problem?

Thank you for your fast responce!

Kind regards
Sergio

 

[tom]

Re: I/O Errors: Supermicro 1027R-N3RF, Samsung SSD 840 Pro MZ-7PD512BW, Areca ARC-168

as a general recommendation, if you use SSDs or HDDs on a raid controller you need to make sure that these SSD or hard drives are on the compatibility list.

areca publishes such a list for SSD here:
http://www.areca.us//support/download/RaidCards/Documents/Hardware/SSDCompatibilityList.zip

 

[udo]

Re: I/O Errors: Supermicro 1027R-N3RF, Samsung SSD 840 Pro MZ-7PD512BW, Areca ARC-168

Hi

I've got some big problems with my proxmox setup.

Generally all works fine now, at the beginning i got some more problems with software-raid setup. Now I've changed that to an Areca ARC-1680 hardware RAID-controller with 4 Intel SSDs in a RAID 5 setup, which seems to work fine...for the first.. Everything is running on a Supermicro 1027R-N3RF Server.

Now there is a problem when copying some big files/many files at same time, i.e. when cloning a machine.

In /var/log/messages i get errors like the following

[INDENT]kernel: arcmsr0: abort device command of scsi id = 0 lun = 0 
kernel: arcmsr0: wait 'abort all outstanding command' timeout 
kernel: arcmsr0: executing hw bus reset .....
kernel: Areca RAID Controller0: F/W V1.51 2012-07-04 & Model ARC-1680
kernel: arcmsr0: abort device command of scsi id = 0 lun = 0
[/INDENT]

Proxmox then is completely unuasable. I/O-Load is too high to do something other..every other thingts u wanna do in that moment doesnt work, not even a ssh-connection i.e.

Does anyone has some experience with this kind of problem?

Thank you for your fast responce!

Kind regards
Sergio

Hi,
how looks the smart-info?

Use the cli-tool cli64 or the web-frontend of the controller.

Any hints in the log?

cli64
CLI> disk info drv=10
Drive Information 
===============================================================
Device Type                        : SATA(5001B4D710A6C01D)
Device Location                    : Enclosure#2 SLOT 02
Model Name                         : INTEL SSDSA2BZ200G3                     
Serial Number                      : CVLV1355006R200
Firmware Rev.                      : 6PB10362
Disk Capacity                      : 200.0GB
Device State                       : NORMAL
Timeout Count                      : 0
Media Error Count                  : 0
Device Temperature                 : 16 C
SMART Read Error Rate              : 255(255)
SMART Spinup Time                  : 100(0)
SMART Reallocation Count           : 100(0)
SMART Seek Error Rate              : 255(255)
SMART Spinup Retries               : 255(255)
SMART Calibration Retries          : N.A.(N.A.)
===============================================================
GuiErrMsg<0x00>: Success.

Udo