Обсуждение: [HACKERS] [PATCH] New command to monitor progression of long running queries

Hi!

On 17.04.2017 15:09, Remi Colinet wrote:

Hello,

I've implemented a new command named PROGRESS to monitor progression of
long running SQL queries in a backend process.


Use case
=======

A use case is shown in the below example based on a table named t_10m
with 10 millions rows.

The table has been created with :

CREATE TABLE T_10M ( id integer, md5 text);
INSERT INTO T_10M SELECT generate_series(1,10000000) AS id,
md5(random()::text) AS md5;

1/ Start a first psql session to run long SQL queries:

[pgadm@rco ~]$ psql -A -d test
psql (10devel)
Type "help" for help.
test=#

The option -A is used to allow rows to be output without formatting work.

Redirect output to a file in order to let the query run without terminal
interaction:
test=# \o out

Start a long running query:
test=# select * from t_10M order by md5;

2/ In a second psql session, list the backend pid and their SQL query

[pgadm@rco ~]$ psql -d test
psql (10devel)
Type "help" for help.

test=# select pid, query from pg_stat_activity ;
  pid  |                   query
-------+-------------------------------------------
 19081 |
 19084 |
 19339 | select pid, query from pg_stat_activity ;
 19341 | select * from t_10m order by md5;
 19727 | select * from t_10m order by md5;
 19726 | select * from t_10m order by md5;
 19079 |
 19078 |
 19080 |
(9 rows)

test=#

Chose the pid of the backend running the long SQL query to be monitored.
Above example is a parallel SQL query. Lowest pid is the main backend of
the query.

test=# PROGRESS 19341;
                                       PLAN
PROGRESS
-------------------------------------------------------------------------------------------
 Gather Merge
   ->  Sort=> dumping tuples to tapes
         rows r/w merge 0/0 rows r/w effective 0/2722972 0%
         Sort Key: md5
         ->  Parallel Seq Scan on t_10m => rows 2751606/3954135 69% blks
125938/161222 78%
(5 rows)

test=#

The query of the monitored backend is:
test=# select * from t_10M order by md5;

Because the table has 10 millions of rows, the sort is done on tapes.


Design of the command
=================

The design of the patch/command is:
- the user issue the "PROGRESS pid" command from a psql session. The pid
is the one of the backend which runs the SQL query for which we want to
get a progression report. It can be determined from the view
pg_stat_activity.
- the monitoring backend, upon receiving the "PROGRESS pid" command from
psql utility used in step above, sends a signal to the backend whose
process pid is the one provided in the PROGRESS command.
- the monitored backend receives the signal and notes the request as for
any interrupt. Then, it continues its execution of its SQL query until
interrupts can be serviced.
- when the monitored process can service the interrupts, it deals with
the progress request by collecting its execution tree with the execution
progress of each long running node. At this time, the SQL query is no
more running. The progression of each node is calculated during the
execution of the SQL query which is at this moment stopped. The
execution tree is dumped in shared memory pages allocated at the start
of the server. Then, the monitored backend set a latch on which the
monitoring process is waiting for. It then continues executing its SQL
query.
- the monitoring backend collects the share memory data dumped by the
monitored backed, and sends it to its psql session, as a list of rows.

The command PROGRESS does not incur any slowness on the running query
because the execution progress is only computed upon receiving the
progress request which is supposed to be seldom used.

The code heavily reuses the one of the explain command. In order to
share as much code as possible with the EXPLAIN command, part of the
EXPLAIN code which deals with reporting quals for instance, has been
moved to a new report.c file in the src/backend/commands folder. This
code in report.c is shared between explain.c source code and PROGRESS
command source code which is in progress.c file.

The progression reported by PROGRESS command is given in terms of rows,
blocks, bytes and percents. The values displayed depend on the node type
in the execution plan.

The current patch implements all the possible nodes which could take a
lot of time:
- Sequential scan nodes with rows and block progress (node type
T_SeqScan, T_SampleScan, T_BitmapHeaepScan, T_SubqueryScan,
T_FunctionScan, T_ValuesScan, T_CteScan, T_WorkTableScan)
- Tuple id scan node with rows and blocks progress (T_TidScan)
- Limit node with rows progress (T_Limit)
- Foreign and custom scan with rows and blocks progress (T_ForeignScan,
T_CustomScan)
- Index scan, index only scan and bitmap index scan with rows and blocks
progress

Use cases
========

Some further examples of use are shown below in the test_v1.txt file.


What do you make of this idea/patch?

Does it make sense?

Any suggestion is welcome.

The current patch is still work in progress. It is meanwhile stable. It
can be used with regular queries. Utilities commands are not supported
for the moment.
Documentation is not yet written.

Regards
Remi

I had implemented analogous feature as extension *pg_query_state* [1] the idea of which I proposed in the thread [2]. Together with this extension I provided some patches to postgres core to enable to send custom signals to working backend (similar to your PROCSIG_PROGRESS) and to print the current query state through patches in 'ExplainNode' function.

I had implemented the same mechanics as you:
1) interrupt the working backend through ProcSignal;
2) handle progress request in the CHECK_FOR_INTERRUPTS entry;
3) transfer query state through shared memory to caller.
But criticism of my approach was that the structure 'QueryDesc' on basis of which query state is formed can be inconsistent in the places where CHECK_FOR_INTERRUPTS appears [3].

I plan to propose the custom_signal patch to community as soon as possible and as consequence release *pg_query_state* from dependency on patches to postgres core. In perspective, I want to resolve the problem related to binding to the CHECK_FOR_INTERRUPTS entries perhaps through patching the executor and implement the robust PROGRESS command.


1. https://github.com/postgrespro/pg_query_state
2. https://www.postgresql.org/message-id/dbfb1a42-ee58-88fd-8d77-550498f52bc5%40postgrespro.ru
3. https://www.postgresql.org/message-id/24182.1472745492%40sss.pgh.pa.us


--
Maksim Milyutin
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company

Maksim,

2017-04-18 20:31 GMT+02:00 Maksim Milyutin <m.milyutin@postgrespro.ru>:

On 18.04.2017 17:39, Remi Colinet wrote:

Hello Maksim,

The core implementation I suggested for the new PROGRESS command uses
different functions from the one used by EXPLAIN for its core
implementation.
Some source code is shared with EXPLAIN command. But this shared code is
only related to quals, properties, children, subPlans and few other nodes.

All other code for PROGRESS is new code.

I don't believe explain.c code can be fully shared with the one of the
new PROGRESS command. These 2 commands have different purposes.
The core code used for the new PROGRESS command is very different from
the core code used for EXPLAIN.

Perhaps you will be forced to duplicate significant snippets of functionality from explain.c into your progress.c.

 

Currently, few code is duplicated between EXPLAIN and PROGRESS commands. The duplicated code could be moved to file src/backend/commands/report.c which is used to gather shared code between the 2 commands. I will try to complete this code sharing as much as possible.

The main point is that PROGRESS uses the same design pattern as EXPLAIN by parsing the query tree. The work horse of the PROGRESS command is ProgressNode() which calls recursively sub nodes until we reach leaf nodes such as SeqScan, IndexScan, TupleStore, Sort, Material, ... . EXPLAIN command uses a similar work horse with function ExplainNode() which eventually calls different leaf nodes.

Some of the leaf nodes which are common to the 2 commands have been put in the file src/backend/commands/report.c. May be some further code sharing is also possible for the work horse by using a template function which would call EXPLAIN specific leaf node functions or PROGRESS specific leaf node functions.
 

Regarding the queryDesc state of SQL query upon receiving a request to
report its execution progress, it does not bring any issue. The request
is noted when the signal is received by the monitored backend. Then, the
backend continues its execution code path. When interrupts are checked
in the executor code, the request will be dealt.

Yes, interrupts are checked in the CHECK_FOR_INTERRUPTS entries.

When the request is being dealt, the monitored backend will stop its
execution and report the progress of the SQL query. Whatever is the
status of the SQL query, progress.c code checks the status and report
either that the SQL query does not have a valid status, or otherwise the
current execution state of the SQL query.

SQL query status checking is about:
- idle transaction
- out of transaction status
- null planned statement
- utility command
- self monitoring

Other tests can be added if needed to exclude some SQL query state. Such
checking is done in void HandleProgressRequest(void).
I do not see why a SQL query progression would not be possible in this
context. Even when the queryDescc is NULL, we can just report a <idle
transaction> output. This is currently the case with the patch suggested.

It's interesting question - how much the active query is in a usable state on the stage of execution. Tom Lane noticed that CHECK_FOR_INTERRUPTS doesn't give us 100% guarantee about full consistency [1].

I wonder what you mean about usable state.

Currently, the code suggested tests the queryDesc pointer and all the sub nodes pointers in order to detect NULL pointers. When the progress report is collected by the backend, this backend does the collect and consequently does not run the query. So the query tree is not being modified. At this moment, whatever is the query state, we can manage to deal with its static state. It is only a tree which could also be just a NULL pointer in the most extreme case. Such case is dealt in the current code.
 

So far, I've found this new command very handy. It allows to evaluate
the time needed to complete a SQL query.

Could you explain how you get the percent of execution for nodes of plan tree and overall for query?

The progress of execution of the query is computed as follows at 2 different places for each leaf node type (Scan, IndexScan, Sort, Material, TupleStore, ...):

- one place in the executor code, or in access methods code, or in sort utilities code, used during the execution of the SQL query in which following values are counted for instance: rows R/W, blocks, R/W, tapes R/W used for sort, tuple store R/W, ... . Some of these values are already computed in the current Postgresql official source code. Some other values had to be added and collected.

- one place in the leaf function of each node type (ProgressScan(), ProgressSort(), ...) in which percents are computed and are then dumped together with raw values collected during execution, in the report. The dump details can be selected with the VERBOSE option of the PROGRESS command (For instance # PROGRESS VERBOSE $pid)

For instance:

1/ read/write rows are collected when running the executor in the file src/backend/executor/execProcnode.c
==============================================================================

This is already implemented in the current official source tree. Nothing mode needs to be implemented to collect values (total rows number and current rows number already fetched are collected).
The report is implemented in leaf function ProgressScanRows().

2/ read/write blocks are collected in the file src/backend/access/heap/heapam.c
==========================================================

This is already implemented in the current official source tree. Nothing more needs to be implemented to collect values during execution.
 The report is implemented in a leaf function ProgressScanBlks().

3/ the sort progressions are collected in the file src/backend/utils/sort/tuplesort.c
==========================================================

[root@rco pg]# git diff --stat master.. src/backend/utils/sort/tuplesort.c
 src/backend/utils/sort/tuplesort.c | 142 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 127 insertions(+), 15 deletions(-)
[root@rco pg]#

New fields have been added to compute the different I/O (read/write/merge/...) per tapes for instance, during a sort on tape.

The report of Sort node is implemented in leaf function ProgressSort()

 
4/ the tuple store progressions are computed in the file src/backend/utils/sort/tuplestore.c
=================================================================

[root@rco pg]# git diff --stat master.. src/backend/utils/sort/tuplestore.c
 src/backend/utils/sort/tuplestore.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++-----------------
 1 file changed, 56 insertions(+), 17 deletions(-)
[root@rco pg]#

New fields have been added to collect the I/O needed for such tuple store.

The report of TupleStore node is implemented in leaf function ProgressTupleStore().

Other node types have been implemented: TidScan, IndexScan, LimitScan, CustomScan, Hash, ModifyTable.

Such node may require some new fields to collect values during the SQL query execution.

Overall, the overhead caused by new values collected during the SQL query execution, is very low.

A few values need to be collected.

A further improvement would be to report the memory, disk and time
resources used by the monitored backend. An overuse of memory, disk and
time resources can prevent the SQL query to complete.

This functionality is somehow implemented in explain.c. You can see my patch to this file [2]. I only manipulate runtime statistics (data in the structure 'Instrumentation') to achieve the printing state of running query.

I will check your patch and try to add such feature to the current patch.
It provides a valuable hint to estimate whether a SQL query has a chance to complete and will not reached the resource limits.
.

1. https://www.postgresql.org/message-id/24182.1472745492%40sss.pgh.pa.us
2. https://github.com/postgrespro/pg_query_state/blob/master/runtime_explain.patch

Thanks for your suggestions and comments

Regards

Remi

--
Maksim Milyutin
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company

On Mon, Apr 17, 2017 at 9:09 PM, Remi Colinet <remi.colinet@gmail.com> wrote:

Hello,

I've implemented a new command named PROGRESS to monitor progression of long running SQL queries in a backend process.

Thank you for the patch.

I am testing your patch but after applying your patch other regression test failed.

$ make installcheck

13 of 178 tests failed.

Regards,

Vinayak 

On 5 May 2017 at 22:38, Vinayak Pokale <vinpokale@gmail.com> wrote:
>
> On Mon, Apr 17, 2017 at 9:09 PM, Remi Colinet <remi.colinet@gmail.com>
> wrote:
>>
>> Hello,
>>
>> I've implemented a new command named PROGRESS to monitor progression of
>> long running SQL queries in a backend process.
>>
> Thank you for the patch.
>

sorry if i'm bikeshedding to soon but... why a command instead of a function?
something like pg_progress_backend() will be in sync with
pg_cancel_backend()/pg_terminate_backend() and the result of such a
function could be usable by a tool to examine a slow query status

--
Jaime Casanova                      www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services