Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-07-02 Thread Kevin Grittner
In preparing to push the patch, I noticed I hadn't responded to this: Gurjeet Singh gurj...@singh.im wrote: Kevin Grittner kgri...@ymail.com wrote: I have reviewed this patch, and think we should do what the patch is trying to do, but I don't think the submitted patch would actually work.

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-07-02 Thread Kevin Grittner
Gurjeet Singh gurj...@singh.im wrote: On Wed, Jul 2, 2014 at 3:49 PM, Kevin Grittner kgri...@ymail.com wrote: Gurjeet Singh gurj...@singh.im wrote: Kevin Grittner kgri...@ymail.com wrote: I have reviewed this patch, and think we should do what the patch is trying to do, but I don't think the

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-07-01 Thread Kevin Grittner
Tom Lane t...@sss.pgh.pa.us wrote: If we're going to do it like this, then I think the force flag should be considered to do nothing except override the clock check, which probably means it shouldn't be tested in the initial if() at all. That makes sense, and is easily done.  The only

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-06-29 Thread Kevin Grittner
I have reviewed this patch, and think we should do what the patch is trying to do, but I don't think the submitted patch would actually work.  I have attached a suggested patch which I think would work.  Gurjeet, could you take a look at it? My comments on prior discussion embedded below.

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-06-29 Thread Tom Lane
Kevin Grittner kgri...@ymail.com writes: Gurjeet Singh gurj...@singh.im wrote: Besides, there's already a throttle built in using the PGSTAT_STAT_INTERVAL limit. This is a key point; neither the original patch nor my proposed alternative bypasses that throttling. That's a fair point that

[HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Gurjeet Singh
Please find attached the patch to send transaction commit/rollback stats to stats collector unconditionally. Without this patch, the transactions that do not read from/write to a database table do not get reported to the stats collector until the client disconnects. Hence the transaction counts

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Tom Lane
Gurjeet Singh gurj...@singh.im writes: Please find attached the patch to send transaction commit/rollback stats to stats collector unconditionally. That's intentional to reduce stats traffic. What kind of performance penalty does this patch impose? If the number of such transactions is large

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Gurjeet Singh
On Wed, Mar 19, 2014 at 4:22 PM, Tom Lane t...@sss.pgh.pa.us wrote: Gurjeet Singh gurj...@singh.im writes: Please find attached the patch to send transaction commit/rollback stats to stats collector unconditionally. That's intentional to reduce stats traffic. What kind of performance

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Alvaro Herrera
Gurjeet Singh wrote: On Wed, Mar 19, 2014 at 4:22 PM, Tom Lane t...@sss.pgh.pa.us wrote: Gurjeet Singh gurj...@singh.im writes: Please find attached the patch to send transaction commit/rollback stats to stats collector unconditionally. That's intentional to reduce stats traffic.

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Tom Lane
Alvaro Herrera alvhe...@2ndquadrant.com writes: I'm not sure I understand the point of this whole thing. Realistically, how many transactions are there that do not access any database tables? I think that something like select * from pg_stat_activity might not bump any table-access counters,

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Robert Haas
On Wed, Mar 19, 2014 at 7:27 PM, Tom Lane t...@sss.pgh.pa.us wrote: Alvaro Herrera alvhe...@2ndquadrant.com writes: I'm not sure I understand the point of this whole thing. Realistically, how many transactions are there that do not access any database tables? I think that something like

Re: [HACKERS] Patch to send transaction commit/rollback stats to the stats collector unconditionally.

2014-03-19 Thread Gurjeet Singh
On Wed, Mar 19, 2014 at 7:27 PM, Tom Lane t...@sss.pgh.pa.us wrote: Alvaro Herrera alvhe...@2ndquadrant.com writes: I'm not sure I understand the point of this whole thing. Realistically, how many transactions are there that do not access any database tables? I think that something like