Horton

Re: Intermittent buildfarm failures on wrasse

Alvaro Herrera <alvherre@alvh.no-ip.org>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Masahiko Sawada <sawada.mshk@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Andres Freund <andres@anarazel.de>, pgsql-hackers@lists.postgresql.org, Noah Misch <noah@leadboat.com>, Peter Geoghegan <pg@bowt.ie>, David Rowley <dgrowleyml@gmail.com>, Amit Kapila <amit.kapila16@gmail.com>
Date: 2022-05-14T14:53:00Z
Lists: pgsql-hackers

Attachments

On 2022-Apr-20, Masahiko Sawada wrote:

> >     MyProc->statusFlags = (MyProc->statusFlags & ~PROC_XMIN_FLAGS) |
> >                           (proc->statusFlags & PROC_XMIN_FLAGS);
> >
> > Perhaps the latter is more future-proof.

> Copying only xmin-related flags in this way also makes sense to me and
> there is no problem at least for now. A note would be that when we
> introduce a new flag that needs to be copied in the future, we need to
> make sure to add it to PROC_XMIN_FLAGS so it is copied. Otherwise a
> similar issue we fixed by 0f0cfb494004befb0f6e could happen again.

OK, done this way -- patch attached.

Reading the comment I wrote about it, I wonder if flags
PROC_AFFECTS_ALL_HORIZONS and PROC_IN_LOGICAL_DECODING should also be
included.  I think the only reason we don't care at this point is that
walsenders (logical or otherwise) do not engage in snapshot copying.
But if we were to implement usage of parallel workers sharing a common
snapshot to do table sync in parallel, then it ISTM it would be
important to copy at least the latter.  Not sure there are any cases
were we might care about the former.

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/
"Every machine is a smoke machine if you operate it wrong enough."
https://twitter.com/libseybieda/status/1541673325781196801

Commits

  1. Repurpose PROC_COPYABLE_FLAGS as PROC_XMIN_FLAGS

  2. Tighten ComputeXidHorizons' handling of walsenders.

  3. Adjust VACUUM's removable cutoff log message.

  4. Temporarily add some probes of tenk1's relallvisible in create_index.sql.

  5. Set synchronous_commit=on in test_setup.sql.

  6. Rearrange core regression tests to reduce cross-script dependencies.

  7. snapshot scalability: Don't compute global horizons while building snapshots.