[ovirt-devel] uuid_generate_v1 not random enough (was: Change in ovirt-engine[ovirt-engine-4.0.2]: build: ovirt-engine-4.0.2.3)

[Eli Mesika]

Generating the UUIDs outside the database is not possible always , for
example it is not possible when you install the DBV with initial data ....
so , in any case we need this support inside the database engine

On Mon, Aug 1, 2016 at 4:16 PM, Juan Hernández <jhernand at redhat.com> wrote:

> On 08/01/2016 11:02 AM, Yedidyah Bar David wrote:
> > On Sun, Jul 31, 2016 at 6:35 PM, Jenkins CI <gerrit2 at gerrit.ovirt.org>
> wrote:
> >> Jenkins CI has posted comments on this change.
> >>
> >> Change subject: build: ovirt-engine-4.0.2.3
> >> ......................................................................
> >>
> >>
> >> Patch Set 2:
> >>
> >> Build Failed
> >>
> >>
> http://jenkins.ovirt.org/job/ovirt-engine_4.0.2_check-merged-el7-x86_64/64/
> : FAILURE
> >
> > 15:07:43 Running upgrade sql script
> >
> './packaging/dbscripts/upgrade/03_05_0580_add_default_instance_types.sql'...
> > 15:07:43
> psql:./packaging/dbscripts/upgrade/03_05_0580_add_default_instance_types.sql:131:
> > ERROR:  duplicate key value violates unique constraint
> > "pk_permissions_id"
> > 15:07:43 DETAIL:  Key (id)=() already exists.
> >
> > Didn't try reproducing, but my best guess is that above uuid
> > was generated by a call to uuid_generate_v1() in above file after it
> > was already inserted to the table in
> > dbscripts/data/00600_insert_permissions.sql.
> >
> > Perhaps we should be using something more random.
> >
> > [1] https://www.postgresql.org/docs/9.5/static/uuid-ossp.html
> >
>
> Note that we did use that long time ago, and we decided to stop using it
> because enabling that extension required an additional RPM package
> (postgrersql-contrib) and different mechanisms to enable it in
> PostgreSQL 8 and PostgreSQL 9, which was cumbersome. As we should now
> support only PostgreSQL 9.2 or newer, this may be no longer a problem.
> In that case you can just remove our "uuid_generate_v1" function and
> enable the extension, the name of the function is the same.
>
> However, it would be better, in general, to generate the identifiers
> outside of the database, and pass them as parameters to the stored
> procedures or SQL statements, that way we have less dependencies.
>
> See bug 870056 for additional information.
>
> >>
> >>
> http://jenkins.ovirt.org/job/ovirt-engine_4.0.2_check-merged-fc23-x86_64/64/
> : SUCCESS
> >>
> >> --
> >> To view, visit https://gerrit.ovirt.org/61733
> >> To unsubscribe, visit https://gerrit.ovirt.org/settings
> >>
> >> Gerrit-MessageType: comment
> >> Gerrit-Change-Id: Ice9518eea22dfa19126e04e49f3a76ee86f3e3c0
> >> Gerrit-PatchSet: 2
> >> Gerrit-Project: ovirt-engine
> >> Gerrit-Branch: ovirt-engine-4.0.2
> >> Gerrit-Owner: Yedidyah Bar David <didi at redhat.com>
> >> Gerrit-Reviewer: Eyal Edri <eedri at redhat.com>
> >> Gerrit-Reviewer: Jenkins CI
> >> Gerrit-Reviewer: Oved Ourfali <oourfali at redhat.com>
> >> Gerrit-Reviewer: Yedidyah Bar David <didi at redhat.com>
> >> Gerrit-Reviewer: gerrit-hooks <automation at ovirt.org>
> >> Gerrit-HasComments: No
> >
> >
> >
>
>
> --
> Dirección Comercial: C/Jose Bardasano Baos, 9, Edif. Gorbea 3, planta
> 3ºD, 28016 Madrid, Spain
> Inscrita en el Reg. Mercantil de Madrid – C.I.F. B82657941 - Red Hat S.L.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/infra/attachments/20160802/a5817990/attachment.html>