[Samba] Problems running samba in vmware

[Scott Lovenberg]

Douglas VanLeuven wrote:
> Adam Zimmer wrote:
>   
>> At the moment I have enabled timeSync with vmware tools.
>>
>> In the general area of time keeping on the host, I added the following
>> settings which avoided errors about the RTC missing interrupts:
>> host.usefastclock=false
>> host.cpukHz=2400000
>> host.useTSC=true
>> ptsc.useTSC=true
>>
>> I have two other machines similarly configured (with the exception of
>> running other linux applications not samba).
>>
>> Ntpdate seems to be installed as it is part of the ubuntu-server default
>> config. However, my other machines seem to run it ok. If anything they
>> fall behind a bit and the vmware sync keeps them up-to-date.
>>
>>     
>
>   
>> Ian McDonald wrote:
>>     
>>> How are your time sync options set for the VM? Is it keeping time ok?
>>> (note,AFAIR, you're not supposed to run NTP within a VM.).
>>>
>>>       
>
> True.  I refer to this document from vmware.
> http://www.vmware.com/pdf/vmware_timekeeping.pdf
>
> Generally, ntp & vmware timesync fight each other.  The usual method is
> to turn off the ntp service, figure out how to minimize interrupts,
> allow the clock to run a little slow and allow vmware timesync to bump
> up the time when it gets about 1 minute slow.
>
> There's another thread that mentions issues with on-board nics and
> drivers.  Over the years, I've bumped into that myself.  To the extent I
>  try and use host-only and route whenever possible.  That's worked
> better for me in generic usage.
>
> Regards, Doug
>   
Just an idea, although I've never tried it in vmware, if you can somehow 
make it a gig network connection and bring up the MTU and even enable 
NAPI in the guest, that should cut down on the IRQs, and slow clock 
drift.  Also, if you have a VMI kernel on the guest (that might be 
VMWare server - 2.0 only, not sure), it should play a little nicer.  
Also, if you can turn off hardware offloading in the guest, it probably 
couldn't hurt. 

With VMs I've found slimmed down kernels really seem to drag less, 
although it could just be the power of suggestion on my own part after 
spending twenty minutes staring at 'make menuconfig'.  Speaking of 
which, if you don't need X, running at runlevel 3 will help, too.