[Ovirt-devel] [PATCH node] Added a wrapper script around the boot menu item to warn the users about the reboot
Perry Myers
pmyers at redhat.com
Wed Jan 14 22:02:17 UTC 2009
Bryan Kearney wrote:
> Perry Myers wrote:
>> Bryan Kearney wrote:
>>> Perry Myers wrote:
>>>> Question on this patch, why did you create a separate script to wrap
>>>> the o-c-boot script instead of just putting a question at the top of
>>>> the main section of the o-c-boot script itself? You could just have
>>>> o-c-boot be executed directly by the firstboot menu and the first
>>>> thing it does is say "Are you sure" and if the user says No then it
>>>> just exits back to the main menu.
>>>
>>> This script is called several times, and it seemed ugly to put in an
>>> "Interactive" parameter to look for. Since it is called interactive
>>> in only one place, I figured it was cleaner to layer as opposed to
>>> add the if checks and flag.
>>
>> Ah, got it. I was only thinking about the interactive case. In the
>> non-interactive case the o-c-boot script is called directly w/o the
>> wrapper. Ok, given that this seems ok to me.
>>
>>>>
>>>> Also, do we have sanity checks in place for o-c-boot that check to
>>>> make sure that other things have been run like o-c-networking and
>>>> o-c-storage? If we don't have this, add that to the top of main in
>>>> o-c-boot along with the user question.
>>>
>>> We do some basic checking, but nothing rock-solid. Do we check for an
>>> initial state, or just provide text saying "You need to have done
>>> X,Y,Z for this to work"
>>
>> The latter won't work because o-c-storage doesn't guarantee that an
>> ovirt Node partition was created, correct? o-c-storage can be run
>> just to create partitions for config/logging but not hard disk boot.
>> So the user could run o-c-storage, elect not to create a partition to
>> store the Node itself and then try to run o-c-boot. In this case
>> o-c-boot would fail.
>>
>> So I think we want to do checks that ensure that the minimum
>> partitions are present with minimum recommended size
> I looked again, and o-c-boot does check for a basic partition structure,
> so I think we are good with that. Just to confirm, the additional check
> would be to see if there is at least network script like
> /etc/sysconfig/network-scripts/ifcfg-ovirtbr*
>
> Is that the correct logic?
Hmm. Perhaps not. I suppose there is a use case for someone using
standalone mode w/o a network at all. In this case you would not be able
to configure a network interface or might not even have one present. But
you would still need a virtual network to put your vm's on so they could
talk to each other.
So I think we just need the check for the partitions (which you say is
already there). We should be all set then.
Perry
More information about the ovirt-devel
mailing list