Sophie Lemaitre wrote:
> Hello,
>
> YAIM states explicitely that the fact that the indexes already exist is harmless.
> And so does the documentation.
>
>
Well, it's in the documentation since yesterday, 15:37.
I've been fighting with the problem since 9 o'clock in the morning. I
wrote my mail afterwards, before refreshing the documentation page. My
fault.
Besides, I didn't see ANYTHING about this weird behavior in the update
16 release notes, that were broadcasted to site admins. There is also
nothing about it on the updates web page (and much less grave torque
issue is explicitly stated in red). There is only enigmatic general
statement "Please also pay attention to the individual known issues that
can be found on the individual service web pages".
Anyways, I think that if doing the migration with running services
corrupts the database, it would be generally a good idea to add
something like "service dpm stop" before the migration if the previous
version is lower than 1.6.0. If the postinstall script is smart enough
to perform the migration if necessary, it could be also kind enough to
stop the services. It would save some people a lot of headache.
Best regards,
Adam
|