On Mon, Apr 19, 2021 at 09:41:30AM -0700, Andres Freund wrote:
> On 2021-04-17 19:13:24 -0500, Justin Pryzby wrote:
> > I'm now realizing that that's RAM use for a single query, not from continuous
> > leaks across multiple queries.
>
> What's the memory usage with inlining disabled, and whats the usage
> without jit?
PGOPTIONS="-c jit=off -c jit_inline_above_cost=0 -c jit_above_cost=0 -c client_min_messages=debug -c
log_executor_stats=on"python3 `pwd`/jitleak.py
! [5.272604 s user, 0.329912 s system total]
! 120948 kB max resident size
PGOPTIONS="-c jit=on -c jit_inline_above_cost=0 -c jit_above_cost=0 -c client_min_messages=debug -c
log_executor_stats=on"python3 `pwd`/jitleak.py
DEBUG: time to inline: 0.003s, opt: 0.000s, emit: 0.003s
... repeated many times ...
! [585.090087 s user, 31.835228 s system total]
! 4639556 kB max resident size
PGOPTIONS="-c jit=on -c jit_inline_above_cost=-1 -c jit_above_cost=0 -c client_min_messages=debug -c
log_executor_stats=on"python3 `pwd`/jitleak.py
DEBUG: time to inline: 0.000s, opt: 0.000s, emit: 0.003s
... repeated many times ... which is confusing, since inlining is disabled ...
! [318.556489 s user, 8.413007 s system total]
! 159436 kB max resident size
> > Maybe we're missing this condition somehow ?
> > if (llvm_jit_context_in_use_count == 0 &&
>
> Do you utilize cursors or such?
Not in this backend/process.
I believe postgis/mapserver uses cursors in a separate process, though...
Our report runs like:
SELECT large_object_oid FROM ... WHERE name=$1 AND user=$2 ... FOR UPDATE -- our reports are stored as LOs and read by
readlo
begin;
SET enable_nestloop=off; -- I'm sorry to say but nothing else has avoids occasional pathological query plans
SELECT report text...;
rollback;
> > Also, I just hit this assertion by cancelling the query with ^C / sigint. But
> > I don't have a reprodcer for it.
> >
> > < 2021-04-17 19:14:23.509 ADT telsasoft >PANIC: LLVMJitContext in use count not 0 at exit (is 1)
>
> Hm, I'll try to see how that happens - if not, do you have a backtrace?
This may be because I backpatched to v13, which had one conflict (0a2bc5d61).
Maybe this is relevant, too: bab150045
If you could provide a patch for v13, it's a good bet there's no issue that I
didn't cause.
I'm able to reproduce the crash like this, with my patch to v13, but not your
patch on master.
python3 -c "import pg; db=pg.DB('dbname=postgres'); db.query('SET jit_above_cost=0; SET jit_inline_above_cost=0; SET
jit=on;SET client_min_messages=debug'); db.query('begin'); db.query_formatted('SELECT 1 FROM generate_series(1,99)a
WHEREa=%s', [1], inline=False);"
--
Justin
