On Mon, Sep 21, 2009 at 03:28:58PM +0100, Michael Hanselmann wrote: > 2009/9/21 Iustin Pop <[email protected]>: > > On Mon, Sep 21, 2009 at 03:21:32PM +0100, Michael Hanselmann wrote: > >> 2009/9/21 Iustin Pop <[email protected]>: > >> > --- a/doc/design-2.1.rst > >> > +++ b/doc/design-2.1.rst > >> > +Introducting persistent UUIDs > >> > >> s/Introducting/Introducing/ > >> > >> > +Upgrading from Ganeti 2.0 will add automatically uuids to all entities > >> > +missing them. > >> > >> For consistency and because it's an acronym, write “UUID” always in > >> uppercase. > > > > Well, I write ‘uuid’ when I refer to the actual object field. I'll add > > quotes, > > is that fine? > > Yes, that's fine.
Here's the interdiff (with a realisation that an FQDN cannot be an UUID, so no issue about namespace conflicts): diff --git a/doc/design-2.1.rst b/doc/design-2.1.rst index 36213d8..cdac90e 100644 --- a/doc/design-2.1.rst +++ b/doc/design-2.1.rst @@ -434,8 +434,8 @@ handle both cases. The default kvm vif script will be changed to do so. (Xen doesn't have a ganeti provided script, so nothing will be done for that hypervisor) -Introducting persistent UUIDs -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Introducing persistent UUIDs +~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Current state and shortcomings ++++++++++++++++++++++++++++++ @@ -458,8 +458,8 @@ cluster information, and via RAPI as well. Note that Ganeti will not provide any way to change this attribute. -Upgrading from Ganeti 2.0 will add automatically uuids to all entities -missing them. +Upgrading from Ganeti 2.0 will automatically add an ‘uuid’ attribute +to all entities missing it. Proposed changes part 2 @@ -469,7 +469,9 @@ In the next release (e.g. 2.2), the tracking of objects will change from the name to the UUID internally, and externally Ganeti will accept both forms of identification; e.g. an RAPI call would be made either against ``/2/instances/foo.bar`` or against -``/2/instances/bb3b2e42…``. +``/2/instances/bb3b2e42…``. Since an FQDN must have at least a dot, +and dots are not valid characters in UUIDs, we will not have namespace +issues. Another change here is that node identification (during cluster operations/queries like master startup, “am I the master?” and @@ -479,9 +481,9 @@ hostname-based scheme. Internal tracking refers to the way the configuration is stored; a DRBD disk of an instance refers to the node name (so that IPs can be changed easily), but this is still a problem for name changes; thus -these will be changed to point to the node name to ease renames. +these will be changed to point to the node UUID to ease renames. -The advantages of this change (after the second round of changes), if +The advantages of this change (after the second round of changes), is that node rename becomes trivial, whereas today node rename would require a complete lock of all instances. -- iustin
