10

I've got a error when system start on 2 VMs.

dm-0: WRITE SAME failed. Manually zeroing.

OS: CentOS 7.1 VMware ESXI 5.5

1-st VM has LSI Logic Parallel type for SCSI controller on HDD Datastorage.

2-nd VM has Paravirtual type for SCSI controller on SSD Datastorage.

Is it normal or not? And what to do with this error?

7

It's likely related to the mass storage being VMware in both cases. Basically, it's saying that it tried to execute a SCSI command called WRITE SAME which is used to write the the same block of data over x number of consecutive disk blocks. It's usually used to zero particular regions on a disk.

If the actual zero'ing operation doesn't fail then the biggest problem here is that the OS has to manually say "ok now write this block of zeros here" etc, etc until the zero'ing operation is done. That may increase bandwidth to and from storage for the duration of the zero'ing operation. That's not a hugely common operation for most workloads so you may just ignore it.

If it is causing a problem for you, you'd have to check with a VMWare resource to figure out what you need to do to get WRITE SAME to be supported by the VMWare disks.

  • 1
    Thks, a lot. find next info: VMware Workstation/VMware ESXi allow to install two different virtual SCSI adapters: BusLogic and LsiLogic. However the default kernel from CentOS-7 does not include the corresponding driver for any of them thus resulting in an unbootable system if you install on a SCSI disk using the defaults for CentOS Linux. If you select 'Red Hat Enterprise Linux' as OS, the paravirtualized SCSI adapter is used, which works. – Максим Чепуков May 27 '15 at 13:58
  • Without context I don't really know how to respond. You can definitely use RHEL7 on ESXi though. We have several test EL7 machines running on our ESX environment right now. – Bratchley May 27 '15 at 14:06
  • @МаксимЧепуков comment works on vmware workstation 14 / Centos 7 64 bit - just shutdown the vm & choose vm=> settings=> options=> general & set it as RH Enterprise 64 bit – Stuart Cardall Nov 13 '17 at 17:39
4

On at least ESXi 6 this is to be expected, is not an error (rather it is informational), usually happens once per filesystem after Linux VM startup and comes about from Linux kernel filesystems (typically ext4 in my experience) probing for "hardware" support of the WRITE SAME feature and then falling back to the usual manual zeroing behaviour when it can't be found. See https://unix.stackexchange.com/a/244150/134856 for further details.

If you ever look through VMkernel logs for datastore VMDKs (i.e. not raw disk mappings or VVols etc) attached to non-ancient Linux VMs you will find that the VSCSI driver does not support the guest VM sending WRITE SAME (I believe my ESXi log mentions it was sent an invalid SCSI opcode 0x41). Nothing to worry about!

(PS: VMware is spelt with capital letters only at the start)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.