On Thu, Sep 7, 2017 at 10:42 AM, Amit Langote
<Langote_Amit_f8@lab.ntt.co.jp> wrote:
> I too tend to think that any users who use this masking facility would
> know to expect to get these failures on upgraded clusters with invalid
> pd_lower in meta pages.
Yes, I don't think that an optimization reducing WAL that impacts all
users should be stopped by a small set of users who use an option for
development purposes.
> (PS: I wonder if it is reasonable to allow configuring the error level
> used when a masking failure occurs? Currently, checkXLogConsistency()
> will abort the process (FATAL))
It definitely is worth it in my opinion, perhaps with an on/off switch
to trigger a warning instead. The reason why we use FATAL now is to
trigger more easily red flags for any potential buildfarm runs: a
startup process facing FATAL takes down the standby.
>> Perhaps we should document this point for wal_consistency_check?
>
> Do you mean permanently under wal_consistency_check parameter
> documentation or in the release notes under incompatibilities for the
> affected index types?
Under the parameter itself, in the spirit of a don't-do-that from an
upgraded instance.
--
Michael