Re: Tuning a query with ORDER BY and LIMIT

2022-06-22 Thread Michael van der Kolff
What do you see when you remove the LIMIT clause? It may be possible to rewrite this using ROW_NUMBER. --Michael On Thu, Jun 23, 2022 at 5:39 AM Dirschel, Steve < steve.dirsc...@thomsonreuters.com> wrote: > I am fairly new to tuning Postgres queries. I have a long background > tuning Oracle

Re: GSSAPI authentication

2022-06-06 Thread Michael van der Kolff
rds, and you might want to capture DNS traffic on the two hosts. Of course, I have no idea whether that is actually the issue. I remember reading these docs ages ago - best of luck! --Michael On Mon, Jun 6, 2022 at 11:42 PM Michael van der Kolff < mvanderko...@gmail.com> wrote: > Oh

Re: GSSAPI authentication

2022-06-06 Thread Michael van der Kolff
Oh wait, I see. On Mon, Jun 6, 2022 at 11:41 PM Michael van der Kolff < mvanderko...@gmail.com> wrote: > The part that you're missing, I think, is that Kerberized services require > a service account. > > The SPN (service principal name) is the name that is used in K

Re: GSSAPI authentication

2022-06-06 Thread Michael van der Kolff
On Mon, Jun 6, 2022 at 11:33 PM Niels Jespersen wrote: > *Fra:* Michael van der Kolff > *Sendt:* 6. juni 2022 14:26 > *Til:* Niels Jespersen > *Cc:* pgsql-general list > *Emne:* Re: GSSAPI authentication > > > > >This sounds like your PG service was unabl

Re: GSSAPI authentication

2022-06-06 Thread Michael van der Kolff
to your question. --Michael On Mon, Jun 6, 2022 at 10:26 PM Michael van der Kolff < mvanderko...@gmail.com> wrote: > This sounds like your PG service was unable to authenticate itself to AD. > > There's probably a trick to that somewhere - AD doesn't really want to be > a Kerber

Re: GSSAPI authentication

2022-06-06 Thread Michael van der Kolff
This sounds like your PG service was unable to authenticate itself to AD. There's probably a trick to that somewhere - AD doesn't really want to be a Kerberos server, it just happens to use it  On Mon, 6 June 2022, 10:05 pm Niels Jespersen, wrote: > Hello all > > > > We are running Postgres

Re: Overriding natural order of query results for a subset

2021-05-29 Thread Michael van der Kolff
Have you considered use of the "nulls last" option in order by ( https://www.postgresql.org/docs/13/queries-order.html)? Alternatively, you could write your own type, with its own ordering primitive  On Sun, 30 May 2021, 12:15 am Laura Smith, < n5d9xq3ti233xiyif...@protonmail.ch> wrote: > Hi >

Re: Modelling versioning in Postgres

2021-05-28 Thread Michael van der Kolff
One thing you could consider is a range type for your "versionTS" field instead of a single point in time. So that would be: CREATE TABLE objects ( objectID uuid, versionID uuid, validRange tsrange, objectData text, ); See https://www.postgresql.org/docs/12.5/rangetypes.html for more