Re: age(xid) on hot standby

Excerpts from Peter Eisentraut's message of mié dic 28 15:04:09 -0300 2011:
> The check_postgres txn_wraparound action[0] runs this query:
>
>    SELECT datname, age(datfrozenxid) AS age FROM pg_database WHERE datallowconn ORDER BY 1, 2
>
> On a hot standby, this fails with:
>
>    ERROR:  cannot assign TransactionIds during recovery
>
> So, a couple of things to wonder about:
>
> Is it unreasonable to check for transaction ID wraparound on a standby?
> It should mirror the situation on the primary, shouldn't it?
>
> Should the age(xid) function do something more useful on a standby,
> e.g., have a custom error message or return null or use the transaction
> ID from the master?

I think we could just have the xid_age call
GetCurrentTransactionIdIfAny, and if that returns InvalidXid, use
ReadNewTransactionId instead.  That xid_age assigns a transaction seems
more of an accident than really intended.

--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support