On 19 Jul 2013, at 15:12, Antonio Terceiro <[email protected]> wrote:

> On Fri, Jul 19, 2013 at 11:44:53AM +0100, Dave Pigott wrote:
>> Hi all,
>> 
>> Well, as happens every time I play with the lava cloud nothing goes quite as 
>> smoothly as it should. Essentially, what's happened, is that I've suspended 
>> the staging instance, then I tried the migration, but that failed because of 
>> an OpenStack network issue. This I fixed, but it still won't migrate for 
>> reasons I don't fully understand.
>> 
>> I then tried un-suspending the staging instance and it's not coming back up. 
>> <<sigh>>
>> 
>> The *easy* path to follow would be to blow it all away and just make a new 
>> staging instance on bare metal, but I'm worried about losing the history on 
>> that machine, so I'm going to endeavour to unknot the issue and get staging 
>> back up again so that I can at least backup up the database so we maintain 
>> the history, then I'll move it to bare metal.
>> 
>> Sorry for the delays, but I can't give a timeframe as yet as to when staging 
>> will be available again.
>> 
>> I'd like to canvas people's opinions on how important the staging history 
>> is. It seems to me that it probably isn't that important, but other 
>> perspectives would be useful.
> 
> Before we have a consistent procedure for using tests on staging as
> prerequisite for updating production, my feeling is that the data on
> staging is not so important. That will change once that procedure is in
> place.

I tend to agree. I've managed to rescue the qemu disk image from staging, and 
have it backed up so we can always look at it at a later time if needed.

I'll do a tear down and new deploy of staging on bare metal on Monday.

Thanks

Dave


_______________________________________________
linaro-validation mailing list
[email protected]
http://lists.linaro.org/mailman/listinfo/linaro-validation

Reply via email to