The ExecutorEnd hook is invoked in a context that could be quite
long-lived, not the executor's own per-query context as I think
we were sort of assuming. Thus, any cruft generated while producing
the EXPLAIN output could accumulate over multiple queries. This can
result in spectacular leakage if log_nested_statements is on, and
even without that I'm surprised nobody complained before.
To fix, just switch into the executor's context so that anything we
allocate will be released when standard_ExecutorEnd frees the executor
state. We might as well nuke the code's retail pfree of the explain
output string, too; that's laughably inadequate to the need.
Japin Li, per report from Jeff Janes. This bug is old, so
back-patch to all supported branches.
Discussion: https://postgr.es/m/CAMkU=1wCVtbeRn0s9gt12KwQ7PLXovbpM8eg25SYocKW3BT4hg@mail.gmail.com
A parallel worker process should not be making any decisions of its
own about whether to auto-explain. If the parent session process
passed down flags asking for instrumentation data, do that, otherwise
not. Trying to enable instrumentation anyway leads to bugs like the
"could not find key N in shm TOC" failure reported in bug #15821
from Christian Hofstaedtler.
We can implement this cheaply by piggybacking on the existing logic
for not doing anything when we've chosen not to sample a statement.
While at it, clean up some tin-eared coding related to the sampling
feature, including an off-by-one error that meant that asking for 1.0
sampling rate didn't actually result in sampling every statement.
Although the specific case reported here only manifested in >= v11,
I believe that related misbehaviors can be demonstrated in any version
that has parallel query; and the off-by-one error is certainly there
back to 9.6 where that feature was added. So back-patch to 9.6.
Discussion: https://postgr.es/m/15821-5eb422e980594075@postgresql.org
Query planning is affected by a number of configuration options, and it
may be crucial to know which of those options were set to non-default
values. With this patch you can say EXPLAIN (SETTINGS ON) to include
that information in the query plan. Only options affecting planning,
with values different from the built-in default are printed.
This patch also adds auto_explain.log_settings option, providing the
same capability in auto_explain module.
Author: Tomas Vondra
Reviewed-by: Rafia Sabih, John Naylor
Discussion: https://postgr.es/m/e1791b4c-df9c-be02-edc5-7c8874944be0@2ndquadrant.com
I (Andres) was more than a bit hasty in committing 33001fd7a7
after last minute changes, leading to a number of problems (jit output
was only shown for JIT in parallel workers, and just EXPLAIN without
ANALYZE didn't work). Lukas luckily found these issues quickly.
Instead of combining instrumentation in in standard_ExecutorEnd(), do
so on demand in the new ExplainPrintJITSummary().
Also update a documentation example of the JIT output, changed in
52050ad8eb.
Author: Lukas Fittl, with minor changes by me
Discussion: https://postgr.es/m/CAP53PkxmgJht69pabxBXJBM+0oc6kf3KHMborLP7H2ouJ0CCtQ@mail.gmail.com
Backpatch: 11, where JIT compilation was introduced
Previously, when using parallel query, EXPLAIN (ANALYZE)'s JIT
compilation timings did not include the overhead from doing so on the
workers. Fix that.
We do so by simply aggregating the cost of doing JIT compilation on
workers and the leader together. Arguably that's not quite accurate,
because the total time spend doing so is spent in parallel - but it's
hard to do much better. For additional detail, when VERBOSE is
specified, the stats for workers are displayed separately.
Author: Amit Khandekar and Andres Freund
Discussion: https://postgr.es/m/CAJ3gD9eLrz51RK_gTkod+71iDcjpB_N8eC6vU2AW-VicsAERpQ@mail.gmail.com
Backpatch: 11-
The previous limit of INT_MAX / 1000 seems to have been cargo-culted in
from somewhere else. Or possibly the value was converted to microseconds
at some point; but in all supported releases, it's just compared to other
values, so there's no need for the restriction. This change raises the
effective limit from ~35 minutes to ~24 days, which conceivably is useful
to somebody, and anyway it's more consistent with the range of the core
log_min_duration_statement GUC.
Per complaint from Kevin Bloch. Back-patch to all supported releases.
Discussion: https://postgr.es/m/8ea82d7e-cb78-8e05-0629-73aa14d2a0ca@codingthat.com
Don't move parenthesized lines to the left, even if that means they
flow past the right margin.
By default, BSD indent lines up statement continuation lines that are
within parentheses so that they start just to the right of the preceding
left parenthesis. However, traditionally, if that resulted in the
continuation line extending to the right of the desired right margin,
then indent would push it left just far enough to not overrun the margin,
if it could do so without making the continuation line start to the left of
the current statement indent. That makes for a weird mix of indentations
unless one has been completely rigid about never violating the 80-column
limit.
This behavior has been pretty universally panned by Postgres developers.
Hence, disable it with indent's new -lpl switch, so that parenthesized
lines are always lined up with the preceding left paren.
This patch is much less interesting than the first round of indent
changes, but also bulkier, so I thought it best to separate the effects.
Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org
Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
Previously, it was unsafe to execute a plan in parallel if
ExecutorRun() might be called with a non-zero row count. However,
it's quite easy to fix things up so that we can support that case,
provided that it is known that we will never call ExecutorRun() a
second time for the same QueryDesc. Add infrastructure to signal
this, and cross-checks to make sure that a caller who claims this is
true doesn't later reneg.
While that pattern never happens with queries received directly from a
client -- there's no way to know whether multiple Execute messages
will be sent unless the first one requests all the rows -- it's pretty
common for queries originating from procedural languages, which often
limit the result to a single tuple or to a user-specified number of
tuples.
This commit doesn't actually enable parallelism in any additional
cases, because currently none of the places that would be able to
benefit from this infrastructure pass CURSOR_OPT_PARALLEL_OK in the
first place, but it makes it much more palatable to pass
CURSOR_OPT_PARALLEL_OK in places where we currently don't, because it
eliminates some cases where we'd end up having to run the parallel
plan serially.
Patch by me, based on some ideas from Rafia Sabih and corrected by
Rafia Sabih based on feedback from Dilip Kumar and myself.
Discussion: http://postgr.es/m/CA+TgmobXEhvHbJtWDuPZM9bVSLiTj-kShxQJ2uM5GPDze9fRYA@mail.gmail.com
This patch widens SPI_processed, EState's es_processed field, PortalData's
portalPos field, FuncCallContext's call_cntr and max_calls fields,
ExecutorRun's count argument, PortalRunFetch's result, and the max number
of rows in a SPITupleTable to uint64, and deals with (I hope) all the
ensuing fallout. Some of these values were declared uint32 before, and
others "long".
I also removed PortalData's posOverflow field, since that logic seems
pretty useless given that portalPos is now always 64 bits.
The user-visible results are that command tags for SELECT etc will
correctly report tuple counts larger than 4G, as will plpgsql's GET
GET DIAGNOSTICS ... ROW_COUNT command. Queries processing more tuples
than that are still not exactly the norm, but they're becoming more
common.
Most values associated with FETCH/MOVE distances, such as PortalRun's count
argument and the count argument of most SPI functions that have one, remain
declared as "long". It's not clear whether it would be worth promoting
those to int64; but it would definitely be a large dollop of additional
API churn on top of this, and it would only help 32-bit platforms which
seem relatively less likely to see any benefit.
Andreas Scherbaum, reviewed by Christian Ullrich, additional hacking by me
New configuration parameter auto_explain.sample_ratio makes it
possible to log just a fraction of the queries meeting the configured
threshold, to reduce the amount of logging.
Author: Craig Ringer and Julien Rouhaud
Review: Petr Jelinek
The folly of the previous arrangement was just demonstrated: there's no
convenient way to add fields to ExplainState without breaking ABI, even
if callers have no need to touch those fields. Since we might well need
to do that again someday in back branches, let's change things so that
only explain.c has to have sizeof(ExplainState) compiled into it. This
costs one extra palloc() per EXPLAIN operation, which is surely pretty
negligible.
We've gotten enough push-back on that change to make it clear that it
wasn't an especially good idea to do it like that. Revert plain EXPLAIN
to its previous behavior, but keep the extra output in EXPLAIN ANALYZE.
Per discussion.
Internally, I set this up as a separate flag ExplainState.summary that
controls printing of planning time and execution time. For now it's
just copied from the ANALYZE option, but we could consider exposing it
to users.
This has been broken since commit af7914c662,
which added the EXPLAIN (TIMING) option. Although that commit included
updates to auto_explain, they evidently weren't tested very carefully,
because the code failed to print node timings even when it should, due to
failure to set es.timing in the ExplainState struct. Reported off-list by
Neelakanth Nadgir of Salesforce.
In passing, clean up the documentation for auto_explain's options a
little bit, including re-ordering them into what seems to me a more
logical order.
Prominent binaries already had this metadata. A handful of minor
binaries, such as pg_regress.exe, still lack it; efforts to eliminate
such exceptions are welcome.
Michael Paquier, reviewed by MauMau.
Sometimes it may be useful to get actual row counts out of EXPLAIN
(ANALYZE) without paying the cost of timing every node entry/exit.
With this patch, you can say EXPLAIN (ANALYZE, TIMING OFF) to get that.
Tomas Vondra, reviewed by Eric Theise, with minor doc changes by me.
walsender.h should depend on xlog.h, not vice versa. (Actually, the
inclusion was circular until a couple hours ago, which was even sillier;
but Bruce broke it in the expedient rather than logically correct
direction.) Because of that poor decision, plus blind application of
pgrminclude, we had a situation where half the system was depending on
xlog.h to include such unrelated stuff as array.h and guc.h. Clean up
the header inclusion, and manually revert a lot of what pgrminclude had
done so things build again.
This episode reinforces my feeling that pgrminclude should not be run
without adult supervision. Inclusion changes in header files in particular
need to be reviewed with great care. More generally, it'd be good if we
had a clearer notion of module layering to dictate which headers can sanely
include which others ... but that's a big task for another day.
The previous functions of assign hooks are now split between check hooks
and assign hooks, where the former can fail but the latter shouldn't.
Aside from being conceptually clearer, this approach exposes the
"canonicalized" form of the variable value to guc.c without having to do
an actual assignment. And that lets us fix the problem recently noted by
Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus
log messages about "parameter "wal_buffers" cannot be changed without
restarting the server". There may be some speed advantage too, because
this design lets hook functions avoid re-parsing variable values when
restoring a previous state after a rollback (they can store a pre-parsed
representation of the value instead). This patch also resolves a
longstanding annoyance about custom error messages from variable assign
hooks: they should modify, not appear separately from, guc.c's own message
about "invalid parameter value".
The originally committed patch for modifying CTEs didn't interact well
with EXPLAIN, as noted by myself, and also had corner-case problems with
triggers, as noted by Dean Rasheed. Those problems show it is really not
practical for ExecutorEnd to call any user-defined code; so split the
cleanup duties out into a new function ExecutorFinish, which must be called
between the last ExecutorRun call and ExecutorEnd. Some Asserts have been
added to these functions to help verify correct usage.
It is no longer necessary for callers of the executor to call
AfterTriggerBeginQuery/AfterTriggerEndQuery for themselves, as this is now
done by ExecutorStart/ExecutorFinish respectively. If you really need to
suppress that and do it for yourself, pass EXEC_FLAG_SKIP_TRIGGERS to
ExecutorStart.
Also, refactor portal commit processing to allow for the possibility that
PortalDrop will invoke user-defined code. I think this is not actually
necessary just yet, since the portal-execution-strategy logic forces any
non-pure-SELECT query to be run to completion before we will consider
committing. But it seems like good future-proofing.
This patch also removes buffer-usage statistics from the track_counts
output, since this (or the global server statistics) is deemed to be a better
interface to this information.
Itagaki Takahiro, reviewed by Euler Taveira de Oliveira.
Without these functions, anyone outside of explain.c can't actually use
ExplainPrintPlan, because the ExplainState won't be initialized properly.
The user-visible result of this was a crash when using auto_explain with
the JSON output format.
Report by Euler Taveira de Oliveira. Analysis by Tom Lane. Patch by me.
The original syntax made it difficult to add options without making them
into reserved words. This change parenthesizes the options to avoid that
problem, and makes provision for an explicit (and perhaps non-Boolean)
value for each option. The original syntax is still supported, but only
for the two original options ANALYZE and VERBOSE.
As a test case, add a COSTS option that can suppress the planner cost
estimates. This may be useful for including EXPLAIN output in the regression
tests, which are otherwise unable to cope with cross-platform variations in
cost estimates.
Robert Haas
