Re: monitoring CREATE INDEX [CONCURRENTLY]

Поиск
Список
Период
Сортировка
От Rahila Syed
Тема Re: monitoring CREATE INDEX [CONCURRENTLY]
Дата
Msg-id CAOajBXTE03eXviZSZ=zL-pdkUNsZjGO9pNMi0k2=eF5hMvmLrQ@mail.gmail.com
обсуждение исходный текст
Ответ на Re: monitoring CREATE INDEX [CONCURRENTLY]  (Alvaro Herrera <alvherre@2ndquadrant.com>)
Ответы Re: monitoring CREATE INDEX [CONCURRENTLY]
Список pgsql-hackers
Дерево обсуждения
Hi Alvaro,

Please see few comments below:

1. Makecheck fails currently as view definition of expected rules.out does not reflect latest changes in progress metrics numbering.

2. +      <entry>
+       When creating an index on a partitioned, this column is set to the
+       total number of partitions on which the index is to be created.
+      </entry>
+     </row>
+     <row>
+      <entry><structfield>partitions_done</structfield></entry>
+      <entry><type>bigint</type></entry>
+      <entry>
+       When creating an index on a partitioned, this column is set to the

I think there is a typo here 's/partitioned/partitioned table/'

3.
+       if (hscan->rs_base.rs_parallel != NULL)
+       {
+               ParallelBlockTableScanDesc bpscan;
+
+               bpscan = (ParallelBlockTableScanDesc) hscan->rs_base.rs_parallel;
+               startblock = bpscan->phs_startblock;
+       }
+       else
+               startblock = hscan->rs_startblock;
+
+       /*
+        * Might have wrapped around the end of the relation, if startblock was
+        * not zero.
+        */
+       if (hscan->rs_cblock > startblock)
+               blocks_done = hscan->rs_cblock - startblock;
+       else
+               blocks_done = hscan->rs_nblocks - startblock +
+                       hscan->rs_cblock;
+
+       return blocks_done;

I think parallel scan equivalent bpscan->phs_nblocks along with hscan->rs_nblocks should be used similar to startblock computation above. 

Thank you,
Rahila Syed

On Fri, 29 Mar 2019 at 23:46, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
On 2019-Mar-29, Alvaro Herrera wrote:

> So, CLUSTER and ALTER TABLE rewrites only do non-concurrent index
> builds; and REINDEX can reuse pretty much the same wait-for metrics
> columns as CIC.  So I think it's okay if I move only the metrics that
> conflict for index_build.

The attached version does it that way.  I had to enlarge the param set a
bit more.  (I suspect those extra columns will be useful to reindex.)
Also, rebased for recent conflicting changes.


I think we should consider a new column of an array type, where we could
put things like the list of PIDs to be waited for, the list of OIDs of
index to rebuild, or the list of partitions to build the index on.

--
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


--
Rahila Syed
Performance Engineer
2ndQuadrant 
http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

В списке pgsql-hackers по дате отправления: