Re: BUG #14941: Vacuum crashes

On Fri, Mar 30, 2018 at 05:19:46PM -0700, Andres Freund wrote:
> On 2018-03-30 18:39:01 +0000, Bossart, Nathan wrote:
>> I think the most straightforward approach for fixing this is to add
>> skip-locked functionality in find_all_inheritors(),
>> find_inheritance_children(), and vac_open_indexes(), but I am curious
>> what others think.

For vac_open_indexes() it may make the most sense to introduce a new
try_open_index which wraps on top of try_relation_open with checks on
the opened relation's relkind. 

> I'm actually wondering if we shouldn't just ignore this problem. While
> the other cases VACUUM (SKIP LOCKED) are intended to solve seem common,
> these seem less so. But it'd be a bit weird, too..

Wouldn't that cause deadlocks if one session has VACUUM (SKIP_LOCKED)
and another session running vacuum or autovacuum if simply ignored?
That looks dangerous to me than simply skipping a lock that cannot be
taken.
--
Michael