Hi ,
I have setup a dsm 6.1.7 on esxi 6.7.
The setup is:
- jun loader 1.02b vmware boot option
- loader boot disk is alone on first sata port
- unique data disk is on paravirtualized scsi controller
- network is vmxnet3 10gb
I've install all 6.1.7 updates , only packages installed are stock file station , universal search and open-vm-tools.
The system will be a test environment for a spk package development so I don't need special package but need high i/o speed because my app move a lot of data.
Using 10GBe as network and ssd as vmfs storage I get all performance I want.
Now I'm testing the setup before starting to use it , doing that I found two problem:
First , inspecting logs I see that /var/log/messages has a lot of:
2018-07-05T15:39:57+02:00 syn synostoraged: SmartDataRead(108) read value /dev/sda fail
2018-07-05T15:39:57+02:00 syn synostoraged: disk/disk_temperature_get.c:99 read value /dev/sda fail
This is repeat every minute and the log file continue to grow...
Now , I know that smart is not available in a vmware virtual machine but i would like to know if there's a way to disable smart check and continue error logging.
Second , I've try to enable jumbo frame in the vm to maximize network performance , the virtual switch and the phisical one have both jumbo frame enable ( mtu 9000 ) and so is the client used for test.
I've found a strange problem , when I use vmxnet3 ( only way to get 10gb without assign a direct nic to vm ) i can't raise mtu more than 3050 , I don't know why exactly 3050 , but if I set mtu to 3051 I get a kernel crash and the dsm hangs.
This is the same raising mtu on command line or from dsm interface ( if someone would like to test the best way is use command line because a reboot solve the problem ).
Using e1000e virtual adapter in place of vmxnet3 solve the problem and let me use a mtu of 9000 but obviously I get only 1GB connection.
Is there some place where to see the source of the compiled modules contained in the loader or there's a way to report a bug to who maintain it ?