>>Yes, we can do that. 
>>But only when 1GB hugepages are enabled is this case ? (to not break 
>>migration of current vm with memory hotplug enable) 


Also, the static mem need to be 1024 by socket.

so 2 sockets, 2gb minimum memory  , 4sockets : 4gb minimum memory



----- Mail original -----
De: "aderumier" <[email protected]>
À: "Wolfgang Bumiller" <[email protected]>
Cc: "pve-devel" <[email protected]>
Envoyé: Vendredi 10 Juin 2016 08:51:11
Objet: Re: [pve-devel] [PATCH] add hugepages option v5

> I can add an error like "1GB hugepages are not compatible with memory hotplug 
> currently" if you want. 

>>What if we then start with a $dimm_size of 1G? 
>>At least qemu seems happy with it then? Or is there downside to this? 

Yes, we can do that. 
But only when 1GB hugepages are enabled is this case ? (to not break migration 
of current vm with memory hotplug enable) 


----- Mail original ----- 
De: "Wolfgang Bumiller" <[email protected]> 
À: "aderumier" <[email protected]>, "pve-devel" <[email protected]> 
Envoyé: Mercredi 8 Juin 2016 08:27:08 
Objet: Re: [pve-devel] [PATCH] add hugepages option v5 

> On June 8, 2016 at 5:40 AM Alexandre DERUMIER <[email protected]> wrote: 
> 
> 
> >>When not specifying numaX entries we run into this code using the 
> >>default $dimm_size of 512, which won't work with 1G pages. Iow this 
> >>config without any `numa0`, `numa1`, ... entries: 
> >> hugepages: 1024 
> >> numa: 1 
> >>will error at this point about wrong memory sizes. We could probably 
> >>catch this in foreach_dimm()? 
> 
> About this, this is only when memory hotplug is enabled. 
> It's will not work with 1GB currently, because indeed, we start with 512MB 
> dimm_size. 
> 
> The check is already done in foreach_dimm -> print_mem_object -> 
> hugepages_size 
> 
> 
> I can add an error like "1GB hugepages are not compatible with memory hotplug 
> currently" if you want. 

What if we then start with a $dimm_size of 1G? 
At least qemu seems happy with it then? Or is there downside to this? 
_______________________________________________ 
pve-devel mailing list 
[email protected] 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel 

_______________________________________________
pve-devel mailing list
[email protected]
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Reply via email to