[nova][cinder][ops] question/confirmation of legacy vol attachment migration
mriedemos at gmail.com
Thu Oct 10 18:21:31 UTC 2019
On 10/10/2019 5:00 AM, Gorka Eguileor wrote:
>> 1. Yeah if the existing legacy attachment record doesn't have a connector I
>> was worried about not properly cleaning on for that old connection, which is
>> something I mentioned before, but also as mentioned we potentially have that
>> case when a server is deleted and we can't get to the compute host to get
>> the host connector, right?
> Not really... In that case we still have the BDM info in the DB, so we
> can just make the 3 Cinder REST API calls ourselves (begin_detaching,
> terminate_connection and detach) to have the volume unmapped, the export
> removed, and the volume return to available as usual, without needing to
> go to the storage array manually.
I'm not sure what you mean. Yes we have the BDM in nova but if it's
really old it won't have the host connector stashed away in the
connection_info dict and we won't be able to pass that to the
Are you talking about something else? I realize ^ is very edge case
since we've been storing the connector in the BDM.connection_info since
I think at least Liberty or Mitaka.
>> 2. If I were to use os-terminate_connection, I seem to have a tricky
>> situation on the migration flow because right now I'm doing:
>> a) create new attachment with host connector
>> b) complete new attachment (put the volume back to in-use status)
>> - if this fails I attempt to delete the new attachment
>> c) delete the legacy attachment - I intentionally left this until the end to
>> make sure (a) and (b) were successful.
>> If I change (c) to be os-terminate_connection, will that screw up the
>> accounting on the attachment created in (a)?
>> If I did the terminate_connection first (before creating a new attachment),
>> could that leave a window of time where the volume is shown as not
>> attached/in-use? Maybe not since it's not the begin_detaching/os-detach
>> API...I'm fuzzy on the cinder volume state machine here.
>> Or maybe the flow would become:
>> a) create new attachment with host connector
> This is a good idea in itself, but it's not taking into account weird
> behaviors that some Cinder drivers may have when you call them twice to
> initialize the connection on the same host. Some drivers end up
> creating a different mapping for the volume instead of returning the
> existing one; we've had bugs like this before, and that's why Nova made
> a change in its live instance migration code to not call
> intialize_connection on the source host to get the connection_info for
Huh...I thought attachments in cinder were a dime a dozen and you could
create/delete them as needed, or that was the idea behind the new v3
attachments stuff. It seems to at least be what I remember John Griffith
always saying we should be able to do.
Also if you can't refresh the connection info on the same host then a
change like this:
Which does just that - refreshes the connection info doing reboot and
start instance operations - would break on those volume drivers if I'm
>> b) terminate the connection for the legacy attachment
>> - if this fails, delete the new attachment created in (a)
>> c) complete the new attachment created in (a)
>> - if this fails...?
>> Without digging into the flow of a cold or live migration I want to say
>> that's closer to what we do there, e.g. initialize_connection for the new
>> host, terminate_connection for the old host, complete the new attachment.
> I think any workaround we try to find has a good chance of resulting in
> a good number of bugs.
> In my opinion our options are:
> 1- Completely detach and re-attach the volume
I'd really like to avoid this if possible because it could screw up
running applications and the migration operation itself is threaded out
to not hold up the restart of the compute service. But maybe that's also
true of what I've got written up today though it's closer to what we do
during resize/cold migrate (though those of course involve downtime for
> 2- Write new code in Cinder
> The new code can be either a new action or we can just add a
> microversion to attachment create to also accept "connection_info", and
> when we provide connection_info on the call the method confirms that
> it's a "migration" (the volume is 'in-use' and doesn't have any
> attachments) and it doesn't bother to call the cinder-volume to export
> and map the volume again and simply saves this information in the DB.
If the volume is in-use it would have attachments, so I'm not following
you there. Even if the volume is attached the "legacy" way from a nova
perspective, using os-initialize_connection, there is a volume
attachment record in the cinder DB (I confirmed this in my devstack
testing and the notes are in my patch). It's also precisely the problem
I'm trying to solve which is without deleting the old legacy attachment,
when you delete the server the volume is detached but still shows up in
cinder as in-use because of the orphaned attachment.
> I know the solution it's not "clean/nice/elegant", and I'd rather go
> with option 1, but that would be terrible user experience, so I'll
> settle for a solution that doesn't add much code to Cinder, is simple
> for Nova, and is less likely to result in bugs.
> What do you think?
> PS: In this week's meeting we briefly discussed this topic and agreed to
> continue the conversation here and retake it on next week's meeting.
Thanks for discussing it and getting back to me.
More information about the openstack-discuss