ODBC Session store for Shibboleth SP, error with NameID insert

[Michael A Grady]

On May 21, 2015, at 6:58 PM, Cantor, Scott <cantor.2 at OSU.EDU> wrote:

> On 5/21/15, 7:55 PM, "Michael A Grady" <mgrady at unicon.net> wrote:
>> 
>> Will do. I figured it might be a condition that shouldn't arise if session management/cookies/everything else had worked as expected. Or that it could be a bug to do the search on a set of conditions that go beyond the key fields, and then assume one could insert a new record when that search doesn't return anything.
> 
> It can still arise because a lot of databases don't actually honor SERIALIZABLE transactions. If a select doesn't take a read lock over the table, then there's no way to prevent non-repeatable reads and the insert can still fail if another insert gets in between, which isn't allowed but happens anyway. But I don't know that that's what happened here, and if the only select is qualified by a where clause, that isn't sufficient to prevent a conflict even if there are no races.
> 

Created issue: https://issues.shibboleth.net/jira/browse/SSPCPP-656

(I almost created under the XML Tooling C++ project, but it appeared that most of the previous StorageService-related issues were under the SP C++ project, so I just put it there.)


--
Michael A. Grady
IAM Architect, Unicon, Inc.