Hey Tomasz,
Yep, that's it. It tries to migrate changenum to commit_id, and so you have
one review request with the changenum being migrated, and another that
already has that commit ID. You'll need to null out the field on one of
those. If one is a discarded change, choose that one.
Christian
I can see two rows one with changenum=1701871 and one for
commit_id='1701871'. What I am looking for exactly? (apologies, 0 knowledge
of this product/db here)
On Wednesday, 7 June 2017 20:15:49 UTC+1, Christian Hammond wrote:
>
> Can you look for all review requests with changenum 1701871
Can you look for all review requests with changenum 1701871 (possibly
commit_id with that value --it's a string in this case), see what turns up?
Christian
On Wed, Jun 7, 2017 at 03:37 'Rob Backhurst' via reviewboard <
reviewboard@googlegroups.com> wrote:
> Hi Christian,
>
> This is from our
Hi Christian,
This is from our DBA...
ok, in that case we either have no problematic dupes or we are not sure how
to find them. All dupes mentioned before have same changenum but different
repository_id.. All items returned by executing below code have different
repository_id:
“select
Hi Rob,
Yeah, setting the field to null. There are unique indexes on
(repository_id, commit_id) and (repository_id, changenum). It's fine for
there to be multiple review requests with the same changenum or commit_id
so long as they have different repository_ids.
If there are duplicates, it'd be
Hi Christian,
Thanks for the quick response.
I'm not particularly comfortable with SQL but i'll ask a colleague if he
can help out - i'll let you know.
Thanks
Rob
On Friday, 26 May 2017 09:24:00 UTC+1, Rob Backhurst wrote:
>
> Hi,
>
> We have recently upgraded our ReviewBoard from version
Hi Rob,
It looks like you have multiple review requests with the same commit
ID/change number in the database. We have unique constraints that should
prevent this, but the constraint may have been added after the review
request was created.
What you can do is unset the field (it's going to be
Hi,
We have recently upgraded our ReviewBoard from version 1.7.22 to 2.5.10 -
as well as also upgrading from RHEL 6.x to CentOS 7.3.1611.
Since then we have a review that cannot be interacted with...can't post a
new review or comment, adjust existing comments etc - or even delete the
review.