From: | Robert Creager <robertc(at)spectralogic(dot)com> |
---|---|
To: | Justin Pryzby <pryzby(at)telsasoft(dot)com> |
Cc: | Thomas Munro <thomas(dot)munro(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, "pgsql-performance(at)lists(dot)postgresql(dot)org" <pgsql-performance(at)lists(dot)postgresql(dot)org> |
Subject: | Re: Need help identifying a periodic performance issue. |
Date: | 2021-11-18 00:18:22 |
Message-ID: | 1ADFE43B-5A43-40DA-B039-2D7E5FD82B6C@spectralogic.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
> On Nov 17, 2021, at 4:18 PM, Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
>
> This message originated outside your organization.
>
> On Wed, Nov 17, 2021 at 09:54:14PM +0000, Robert Creager wrote:
> > We are able to move up to Postgres 13.5, in our ports tree, if that would help. We used pg_upgrade to get from 9.6 to 13.3, so that should work fine going instead to 13.5. We’re almost branching/releasing our code, so it’s not a good time, but if it may help with this problem, we’ll deal with it.
>
> To be clear, I have no specfic reason to believe it would help.
I figured as much, and told our gang that also. Was looking through the release notes a little, need to finish with 4 and look at 5.
> But it would be silly to chase down a problem that someone already fixed 10
> months ago (the source of this problem, or something else that comes up).
Yeah, trying to figure out how feasible it is for us to do quickly. Ok, we have approval, and will be working on an upgraded build in the morning.
>
> In fact I suspect it won't help, and there's an issue with your schema, or
> autovacuum, or postgres.
Well, if it’s our schema, 9.6 didn’t care about it, doesn’t mean there wasn’t one though, and I understand that completely.
So, how do I go about capturing more information for the big brains (you guys) to help figure this out? I have all our resources at mine (and hence your) disposal.
>
> Note that since v10, the version scheme uses only two components, and 13.3 to
> 13.5 is a minor release, similar to 9.6.3 to 9.6.5. So you don't need to use
> pg_upgrade - just update the binaries.
Good, but the previous release of our product was at 9.6, so we’re currently using the pg_upgrade to do that, automated (storage appliance). Just talking out loud, that switching to 13.5 shouldn’t cause us any upgrade issues other than figuring out builds and dependencies and some upgrade testing. And a straight re-run with 13.5 on top of the db should work just fine. When I talk out loud, I sometimes catch the stupid things I’m missing, and it allows others to point out the stupid things I’m missing...
Best,
Robert
From | Date | Subject | |
---|---|---|---|
Next Message | Thomas Munro | 2021-11-18 03:39:42 | Re: Need help identifying a periodic performance issue. |
Previous Message | Justin Pryzby | 2021-11-17 23:18:15 | Re: Need help identifying a periodic performance issue. |