Обсуждение: Postgres not using correct indices for views.

Hi,

we have created restricted view for our tables, so that we can allow 
access to non-gdpr relevant data but hide everything else.

For exactly those views, the Query Planner uses the wrong indices, when 
executing exactly the same query, once it takes 0.1 s and on the views 
it takes nearly 18 sec (it does a full table scan, or uses the wrong 
indices).

Do we have to GRANT additional rights? I see it's using some indices, 
just not the correct ones!

Has anyone experienced the same issues? What can we do about that?

Thanks
BR
Thomas

"Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
> we have created restricted view for our tables, so that we can allow 
> access to non-gdpr relevant data but hide everything else.
> For exactly those views, the Query Planner uses the wrong indices, when 
> executing exactly the same query, once it takes 0.1 s and on the views 
> it takes nearly 18 sec (it does a full table scan, or uses the wrong 
> indices).
> Do we have to GRANT additional rights? I see it's using some indices, 
> just not the correct ones!

Does EXPLAIN show reasonable rowcount estimates when you query
directly, but bad ones when you query via the views?

If so, a likely guess is that you're falling foul of the restrictions
added for CVE-2017-7484:

Author: Peter Eisentraut <peter_e@gmx.net>
Branch: master Release: REL_10_BR [e2d4ef8de] 2017-05-08 09:26:32 -0400
Branch: REL9_6_STABLE Release: REL9_6_3 [c33c42362] 2017-05-08 09:18:57 -0400
Branch: REL9_5_STABLE Release: REL9_5_7 [d45cd7c0e] 2017-05-08 09:19:07 -0400
Branch: REL9_4_STABLE Release: REL9_4_12 [3e5ea1f9b] 2017-05-08 09:19:15 -0400
Branch: REL9_3_STABLE Release: REL9_3_17 [4f1b2089a] 2017-05-08 09:19:23 -0400
Branch: REL9_2_STABLE Release: REL9_2_21 [d035c1b97] 2017-05-08 09:19:42 -0400

    Add security checks to selectivity estimation functions
    
    Some selectivity estimation functions run user-supplied operators over
    data obtained from pg_statistic without security checks, which allows
    those operators to leak pg_statistic data without having privileges on
    the underlying tables.  Fix by checking that one of the following is
    satisfied: (1) the user has table or column privileges on the table
    underlying the pg_statistic data, or (2) the function implementing the
    user-supplied operator is leak-proof.  If neither is satisfied, planning
    will proceed as if there are no statistics available.
    
    At least one of these is satisfied in most cases in practice.  The only
    situations that are negatively impacted are user-defined or
    not-leak-proof operators on a security-barrier view.
    
    Reported-by: Robert Haas <robertmhaas@gmail.com>
    Author: Peter Eisentraut <peter_e@gmx.net>
    Author: Tom Lane <tgl@sss.pgh.pa.us>
    
    Security: CVE-2017-7484


However, if you're not on the latest minor releases, you might
find that updating would fix this for you, because of

Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master Release: REL_12_BR [a0905056f] 2019-05-06 11:54:32 +0100
Branch: REL_11_STABLE Release: REL_11_3 [98dad4cd4] 2019-05-06 11:56:37 +0100
Branch: REL_10_STABLE Release: REL_10_8 [ca74e3e0f] 2019-05-06 11:58:32 +0100
Branch: REL9_6_STABLE Release: REL9_6_13 [71185228c] 2019-05-06 12:00:00 +0100
Branch: REL9_5_STABLE Release: REL9_5_17 [01256815a] 2019-05-06 12:01:44 +0100
Branch: REL9_4_STABLE Release: REL9_4_22 [3c0999909] 2019-05-06 12:05:05 +0100

    Use checkAsUser for selectivity estimator checks, if it's set.
    
    In examine_variable() and examine_simple_variable(), when checking the
    user's table and column privileges to determine whether to grant
    access to the pg_statistic data, use checkAsUser for the privilege
    checks, if it's set. This will be the case if we're accessing the
    table via a view, to indicate that we should perform privilege checks
    as the view owner rather than the current user.
    
    This change makes this planner check consistent with the check in the
    executor, so the planner will be able to make use of statistics if the
    table is accessible via the view. This fixes a performance regression
    introduced by commit e2d4ef8de8, which affects queries against
    non-security barrier views in the case where the user doesn't have
    privileges on the underlying table, but the view owner does.
    
    Note that it continues to provide the same safeguards controlling
    access to pg_statistic for direct table access (in which case
    checkAsUser won't be set) and for security barrier views, because of
    the nearby checks on rte->security_barrier and rte->securityQuals.
    
    Back-patch to all supported branches because e2d4ef8de8 was.
    
    Dean Rasheed, reviewed by Jonathan Katz and Stephen Frost.

            regards, tom lane

"Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
> The planner estimates the correct row counts, but still does the wrong 
> planning.

Hm, I'm not exactly convinced.  You show

> Wrong:
>           ->  Hash Join  (cost=359555.11..1849150.95 rows=1496816 
> width=1508) (actual time=1081.081..24251.466 rows=543231 loops=1)
>                 Hash Cond: (p.customer_id = l.customer_id)
>                 Join Filter: ((p.date - '3 days'::interval day) <= 
> l.duedate)
>                 Rows Removed by Join Filter: 596120

> Correct:
>     ->  Nested Loop  (cost=0.87..2961441.25 rows=515233 width=1471)

The join size estimate seems a lot closer to being correct in the
second case, which could lend support to the idea that statistics
aren't being applied in the first case.

However ... it sort of looks like the planner didn't even consider
the second plan shape in the "wrong" case.  If it had, then even
if it costed it 3X more than it did in the "right" case, the second
plan would still have won out by orders of magnitude.  So there's
something else going on.

Can you show the actual query and table and view definitions?

            regards, tom lane

[ re-adding list ]

"Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
> On 9 Aug 2019, at 0:45, Tom Lane wrote:
>> However ... it sort of looks like the planner didn't even consider
>> the second plan shape in the "wrong" case.  If it had, then even
>> if it costed it 3X more than it did in the "right" case, the second
>> plan would still have won out by orders of magnitude.  So there's
>> something else going on.
>> 
>> Can you show the actual query and table and view definitions?

> View definition:
>   SELECT l.id,
>      l.created_at,
>      ...
>      togdpr(l.comment) AS comment,
>      ...
>     FROM loans l;

Ah-hah.  I'd been thinking about permissions on the table and
view, but here's the other moving part: functions in the view.
I bet you were incautious about making this function definition
and allowed togdpr() to be marked volatile --- which it will
be by default.  That inhibits a lot of optimizations.

I'm guessing about what that function does, but if you could
safely mark it stable or even immutable, I bet this view would
behave better.

            regards, tom lane

> [ re-adding list ]
>
> "Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
>> On 9 Aug 2019, at 0:45, Tom Lane wrote:
>>> However ... it sort of looks like the planner didn't even consider
>>> the second plan shape in the "wrong" case.  If it had, then even
>>> if it costed it 3X more than it did in the "right" case, the second
>>> plan would still have won out by orders of magnitude.  So there's
>>> something else going on.
>>>
>>> Can you show the actual query and table and view definitions?
>
>> View definition:
>>   SELECT l.id,
>>      l.created_at,
>>      ...
>>      togdpr(l.comment) AS comment,
>>      ...
>>     FROM loans l;
>
> Ah-hah.  I'd been thinking about permissions on the table and
> view, but here's the other moving part: functions in the view.
> I bet you were incautious about making this function definition
> and allowed togdpr() to be marked volatile --- which it will
> be by default.  That inhibits a lot of optimizations.
>
> I'm guessing about what that function does, but if you could
> safely mark it stable or even immutable, I bet this view would
> behave better.
>
>             regards, tom lane

Yep that was IT! Perfect, thank you soo much!

Why does it inhibit functionalities like using the correct index, if the 
function is only in the select?
Could that still be improved from pg side?

Thanks again!

What a nice catch!

Sent from my iPad

On Aug 10, 2019, at 6:05 AM, Thomas Rosenstein <thomas.rosenstein@creamfinance.com> wrote:

>> [ re-adding list ]
>>
>> "Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
>>>> On 9 Aug 2019, at 0:45, Tom Lane wrote:
>>>> However ... it sort of looks like the planner didn't even consider
>>>> the second plan shape in the "wrong" case.  If it had, then even
>>>> if it costed it 3X more than it did in the "right" case, the second
>>>> plan would still have won out by orders of magnitude.  So there's
>>>> something else going on.
>>>>
>>>> Can you show the actual query and table and view definitions?
>>
>>> View definition:
>>>  SELECT l.id,
>>>     l.created_at,
>>>     ...
>>>     togdpr(l.comment) AS comment,
>>>     ...
>>>    FROM loans l;
>>
>> Ah-hah.  I'd been thinking about permissions on the table and
>> view, but here's the other moving part: functions in the view.
>> I bet you were incautious about making this function definition
>> and allowed togdpr() to be marked volatile --- which it will
>> be by default.  That inhibits a lot of optimizations.
>>
>> I'm guessing about what that function does, but if you could
>> safely mark it stable or even immutable, I bet this view would
>> behave better.
>>
>>            regards, tom lane
>
> Yep that was IT! Perfect, thank you soo much!
>
> Why does it inhibit functionalities like using the correct index, if the function is only in the select?
> Could that still be improved from pg side?
>
> Thanks again!
>
>

"Thomas Rosenstein" <thomas.rosenstein@creamfinance.com> writes:
> On 9 Aug 2019, at 0:45, Tom Lane wrote:
>> I'm guessing about what that function does, but if you could
>> safely mark it stable or even immutable, I bet this view would
>> behave better.

> Yep that was IT! Perfect, thank you soo much!
> Why does it inhibit functionalities like using the correct index, if the 
> function is only in the select?
> Could that still be improved from pg side?

Possibly, but there's a lot of work between here and there, and it's
limited by how much we want to change the semantics around volatile
functions.  The core problem that's breaking this case for you is
that we won't flatten a view (i.e., pull up the sub-SELECT into the
parent query) if its targetlist has volatile functions, for fear
of changing the number of times such functions get invoked.

Now, we're not totally consistent about that anyway --- for example,
the code is willing to push down qual expressions into an un-flattened
sub-SELECT, which could remove rows from the output of the sub-SELECT's
FROM and thereby reduce the number of calls of any volatile functions
in its tlist.  (That particular behavior is very ancient, and I wonder
whether we'd reject it if it were proposed today.)

The thing that's missing to make this better is to be willing to
push down join quals not just restriction quals.  That would require
being able to make "parameterized paths" for subqueries, which is
something that's on the radar screen but nobody's really worked on it.
There are substantial concerns about whether it'd make subquery planning
noticeably more expensive.

            regards, tom lane