Opened 12 years ago
Closed 12 years ago
#103 closed (fixed)
the slice table in gmoc-db2 has unique key uuid, where the unique key should be name + uuid
Reported by: | chaos@bbn.com | Owned by: | chaos@bbn.com |
---|---|---|---|
Priority: | minor | Milestone: | gmoc.py data submission debugging |
Component: | Database | Version: | |
Keywords: | Cc: | ||
Dependencies: |
Description
The slice
table in gmoc-db2 is:
CREATE TABLE `slice` ( `slice_id` int(11) NOT NULL auto_increment, `name` varchar(128) NOT NULL, `last_update` int(10) NOT NULL default '0', `uuid` varchar(128) default NULL, `created` int(11) NOT NULL, `expires` int(11) NOT NULL, PRIMARY KEY (`slice_id`), UNIQUE KEY `uuid` (`uuid`)
However, uuids are not globally unique --- they're only unique at a given entity (SA in this case). The name
field here contains a slice URN, which should be globally unique at a given time, but not necessarily over time. Therefore, for global uniqueness over time, the combination of name
(URN) and uuid
is needed.
A reminder which may be relevant here: when aggregates refer to a slice, they may report only the slice URN and not the UUID, because they may not know it. So any changes to this table should retain the current (GEC14) submission behavior, in which the scripts resolve a stub slice by looking up the name and using the most recent slice with that name.
Change History (3)
comment:1 Changed 12 years ago by
Owner: | changed from somebody to chaos@bbn.com |
---|---|
Status: | new → assigned |
comment:2 Changed 12 years ago by
Status: | assigned → new |
---|
comment:3 Changed 12 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
Here are those two slices:
(2670,'urn:publicid:IDN+pgeni1.gpolab.bbn.com+slice+ecgtest10',1349209473,'f11febbe-0cc7-11e2-adf1-000c29b09582',1349224365,1349227965,''), (2671,'urn:publicid:IDN+pgeni1.gpolab.bbn.com+slice+ecgtest10',1349214938,'d894c5aa-0cd3-11e2-adf1-000c29b09582',1349229478,1349233078,'');
Looks good.
The slice table now has:
I want to verify that two slices with the same URN and different UUIDs are treated correctly. In order to check this, i need to test with a short-duration slice on pgeni1, so that'll take an hour. I'm working on it now, using the slice with URN: