Re: Valgrind failures in Apply Launcher's bgworker_quickdie() exit

From: Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Valgrind failures in Apply Launcher's bgworker_quickdie() exit
Date: 2018-12-14 02:52:23
Message-ID: CAEepm=0=qgmttSsQ5VaeqSitNh37ecdXrOMosDJ_nv5dSPKSHA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, Dec 14, 2018 at 1:15 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> On December 13, 2018 6:01:04 PM PST, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com> writes:
> >> Since libcrypto.so is implicated, Andres asked me off-list if my
> >> changes to random number state initialisation might be linked to
> >> skink's failures beginning 12 or 15 days ago. It appears not, as it
> >> was green for several runs after that commit.
> >> ...
> >> It's Debian unstable, which could be a factor. Bugs in glibc?
>
> Note it's only failing on master...

Here are the commits between the last green run and the first Valgrind failure:

7d4524aed3 Fri Nov 30 22:53:18 2018 UTC Fix tablespace path TAP test
of pg_verify_checksums for msys
9dc1225855 Fri Nov 30 13:20:49 2018 UTC Silence compiler warning
dcfdf56e89 Fri Nov 30 06:20:43 2018 UTC Fix typo.
5c99513975 Fri Nov 30 01:34:45 2018 UTC Fix various checksum check
problems for pg_verify_checksums and base backups
a1c91dd110 Fri Nov 30 01:14:58 2018 UTC Switch pg_verify_checksums
back to a blacklist
d328991578 Thu Nov 29 23:28:10 2018 UTC Document handling of
invalid/ambiguous timestamp input near DST boundaries.
88bdbd3f74 Thu Nov 29 21:42:53 2018 UTC Add log_statement_sample_rate parameter
826eff57c4 Thu Nov 29 20:53:44 2018 UTC Ensure static libraries have
correct mod time even if ranlib messes it up.
68120427f4 Thu Nov 29 13:01:11 2018 UTC doc: Add appendix detailing
some limits of PostgreSQL
44e22647f8 Thu Nov 29 09:00:08 2018 UTC Add pg_partition_tree to
documentation index
431f1599a2 Thu Nov 29 01:31:12 2018 UTC Add support for
NO_INSTALLCHECK in MSVC scripts
2ac180c286 Thu Nov 29 01:14:26 2018 UTC Fix minor typo in dsa.c.
d79fb5d237 Thu Nov 29 00:39:07 2018 UTC Add missing NO_INSTALLCHECK
in commit_ts and test_rls_hooks
4c703369af Thu Nov 29 00:12:19 2018 UTC Fix handling of synchronous
replication for stopping WAL senders
1a990b207b Wed Nov 28 22:42:54 2018 UTC Have BufFileSize() ereport()
on FileSize() failure.

f2cbffc7a6 Wed Nov 28 12:55:54 2018 UTC Only allow one recovery target setting
eae9143d9a Wed Nov 28 12:34:10 2018 UTC C comment: remove extra '*'
0f9cdd7dca Wed Nov 28 01:12:30 2018 UTC Don't set PAM_RHOST for Unix sockets.
f69c959df0 Wed Nov 28 00:43:08 2018 UTC Do not decode TOAST data for
table rewrites
d1ce4ed2d5 Tue Nov 27 23:48:51 2018 UTC Use wildcard to match parens
after CREATE STATISTICS
d67dae036b Tue Nov 27 22:58:10 2018 UTC Don't count zero-filled
buffers as 'read' in EXPLAIN.
471a7af585 Tue Nov 27 20:16:55 2018 UTC Ensure consistent sort order
of large objects in pg_dump.
b238527664 Tue Nov 27 18:07:03 2018 UTC Fix jit compilation bug on wide tables.
f17889b221 Tue Nov 27 14:16:14 2018 UTC Update ssl test certificates and keys
4c8750a9cc Tue Nov 27 07:26:05 2018 UTC Fix ac218aa4f6 to work on
versions before 9.5.
ac218aa4f6 Tue Nov 27 01:00:43 2018 UTC Update pg_upgrade test for
reg* to include regrole and regnamespace.
7a9d6779d9 Tue Nov 27 00:41:29 2018 UTC doc: fix wording for
plpgsql, add "and"
54bb22f66a Mon Nov 26 23:27:34 2018 UTC Fix typo introduced in 578b229718.
12a53c732c Mon Nov 26 22:37:08 2018 UTC Fix pg_upgrade for oid removal.
70d7e507ef Mon Nov 26 22:32:51 2018 UTC Fix translation of special
characters in psql's LaTeX output modes.
95dcb8fc05 Mon Nov 26 20:30:24 2018 UTC Avoid locale-dependent output
in numericlocale check.
67ed3b9d73 Mon Nov 26 20:24:14 2018 UTC Fix sample output for
hash_metapage_info query
aa2ba50c2c Mon Nov 26 20:18:55 2018 UTC Add CSV table output mode in psql.
9a98984f49 Mon Nov 26 17:41:42 2018 UTC Improve regression test
coverage for psql output formats.
a7eece4fc9 Mon Nov 26 17:31:20 2018 UTC Fix breakage of "\pset format latex".
36d442a25a Mon Nov 26 15:38:19 2018 UTC Clarify that cross-row
constraints are unsupported
664f01b613 Mon Nov 26 07:43:19 2018 UTC Revert "Fix typo in
documentation of toast storage"
058ef3a1a8 Mon Nov 26 06:49:23 2018 UTC Fix typo in documentation of
toast storage
1d7dd18686 Mon Nov 26 02:12:11 2018 UTC Revert all new recent changes
to add PGXS options for TAP and isolation
3955cae0c5 Mon Nov 26 01:49:49 2018 UTC Fix regression test handling
of test_decoding with MSVC
b0b1f4183a Mon Nov 26 00:42:21 2018 UTC Disable temporarily TAP tests
for contrib/bloom/
03faa4a8dd Sun Nov 25 23:39:19 2018 UTC Add PGXS options to control
TAP and isolation tests

Maybe you could try reverting f17889b221 (change of key size) on the
superstition that it has something to do with libcrypto.so, but I have
no clue why Apply Launcher in particular would be affected by that.
Otherwise... assuming 2dedf4d9 doesn't show the problem, you could do
the world's slowest bisect 2dedf4d9 -> 7d4524aed3 in ~5.4 steps...

--
Thomas Munro
http://www.enterprisedb.com

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2018-12-14 04:10:25 Re: valgrind issues on Fedora 28
Previous Message Amit Langote 2018-12-14 02:52:03 Re: slight tweaks to documentation about runtime pruning