From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> |
Cc: | Ondřej Jirman <ienieghapheoghaiwida(at)xff(dot)cz>, pgsql-bugs(at)lists(dot)postgresql(dot)org |
Subject: | Re: BUG #16129: Segfault in tts_virtual_materialize in logical replication worker |
Date: | 2019-11-21 23:34:18 |
Message-ID: | 13241.1574379258@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
I wrote:
> Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> writes:
>> FWIW my hunch is the bug is somewhere in this chunk of code from
>> apply_heap_update:
>> oldctx = MemoryContextSwitchTo(GetPerTupleMemoryContext(estate));
>> ExecCopySlot(remoteslot, localslot);
>> slot_modify_cstrings(remoteslot, rel, newtup.values, newtup.changed);
>> MemoryContextSwitchTo(oldctx);
> I imagine the only reason this code has gotten past the valgrind
> animals is that we're not testing cases where non-replaced columns
> in the subscriber table are of pass-by-ref types.
Actually, it doesn't appear to me that we're testing this with
any non-replaced columns at all. The test modifications in the
attached proposed patch add that. For me, the unpatched code
doesn't crash with this test, but the non-replaced column reads
back as empty which is certainly wrong. Valgrind would likely
complain too, but I didn't try it.
regards, tom lane
Attachment | Content-Type | Size |
---|---|---|
fix-bogus-slot-manipulation-in-logical-rep.patch | text/x-diff | 5.1 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | mayur | 2019-11-21 23:35:09 | Re: BUG #16130: planner does not pick unique btree index and goes for seq scan but unsafe hash index works. |
Previous Message | Tom Lane | 2019-11-21 22:37:25 | Re: BUG #16129: Segfault in tts_virtual_materialize in logical replication worker |