Re: recovery error while running any statement

On Thu, Jan 9, 2025 at 12:01 PM yudhi s <learnerdatabase99(at)gmail(dot)com> wrote:

>
> On Thu, Jan 9, 2025 at 10:21 PM Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com>
> wrote:
>
>> On 1/9/25 08:42, yudhi s wrote:
>> > Hello Experts,
>> > It's postgres aurora version 16. While running the ALTER command on
>> any
>> > object we see an error "/Only RowExclusiveLock or less can be acquired
>> > on database objects during recovery/". If I run any DML it gives an
>> > error stating '/cannot execute UPDATE in a read-only transaction/' ,
>> > then I tried setting "/set transaction read-write/" and it erroring out
>> > with "/cannot set transaction read-write mode during recovery/".
>> >
>> > Want to understand , what is the cause of this error and how to fix
>> this?
>>
>> Per:
>>
>>
>> https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Aurora.AuroraPostgreSQL.html
>>
>> Aurora Postgres is not the community edition, you should probably take
>> this up with AWS support.
>>
>>
>> Sure will raise a ticket. I was trying to understand though , if it's
> possible in postgres to have these sudden "recovery errors" possible
> because of some long running DML/DDL killed unexpectedly using the
> "pg_terminate/pg_cancel" command? And if this error appears in community
> postgres and if it stays for a long time what we used to do?
>

Are you *positive* that you're connected to the *primary* node? Because
this is exactly the kind of error you'll see when connected to the
*secondary* node.

Or maybe, for some reason, AWS failed you over to the secondary node,
making it now the primary node, while what you think is the primary node is
really the secondary node.

That's how "real" Postgresql acts, at least. Not how AWS RDS Postgresql
acts (it's all hidden behind a virtual IP that you always connect to), but
I don't know anything about Aurora.

--
Death to <Redacted>, and butter sauce.
Don't boil me, I'm still alive.
<Redacted> lobster!