Comments on: How to estimate query completion time in MySQL http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/ Everything about MySQL Performance Sat, 21 Nov 2009 05:23:57 -0800 http://wordpress.org/?v=2.8.4 hourly 1 By: Kailash Kumar P http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-340009 Kailash Kumar P Mon, 04 Aug 2008 13:59:14 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-340009 Great work Baron. Great work Baron.

]]> By: Mike http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-308813 Mike Wed, 04 Jun 2008 21:13:24 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-308813 Nice work man Nice work man

]]> By: Baron Schwartz http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-287138 Baron Schwartz Thu, 24 Apr 2008 11:57:34 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-287138 I had thought about this patch too. The KILL command is the only place I know where there might be sample code for identifying another thread. I haven't looked at the code yet, but in theory it might not be that hard to do this. I had thought about this patch too. The KILL command is the only place I know where there might be sample code for identifying another thread. I haven’t looked at the code yet, but in theory it might not be that hard to do this.

]]> By: peter http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286947 peter Thu, 24 Apr 2008 03:38:27 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286947 Baron, Too bad this only works on unloaded server. It would be great to get performance counters for different thread in this case it would be much easier to see the progress. May be we should see if it is possible to implement such patch. Baron,

Too bad this only works on unloaded server. It would be great to get performance counters for different thread in this case it would be much easier to see the progress. May be we should see if it is possible to implement such patch.

]]> By: David Holoboff http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286540 David Holoboff Wed, 23 Apr 2008 14:25:45 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286540 Baron, Thank you for sharing this tool - it is very useful... :) I tested this with inserts, updates and deletes and it works very well. I am going to use this request on our job-related servers to estimate finish time. My estimated time after doing several tests on a 100 million row table was within 5 seconds of the actual time for selects, inserts, updates, and deletes (deletes, being the fastest, had a margin of error of no more than one second). Thanks, David Holoboff Baron,

Thank you for sharing this tool – it is very useful… :)

I tested this with inserts, updates and deletes and it works very well. I am going to use this request on our job-related servers to estimate finish time. My estimated time after doing several tests on a 100 million row table was within 5 seconds of the actual time for selects, inserts, updates, and deletes (deletes, being the fastest, had a margin of error of no more than one second).

Thanks,
David Holoboff

]]> By: Baron Schwartz http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286525 Baron Schwartz Wed, 23 Apr 2008 13:48:24 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286525 Yes, those are the "different situations" I mentioned. If you know how many rows will be affected, and you watch the Handler_* counters, you can see how much work has been done. Yes, those are the “different situations” I mentioned. If you know how many rows will be affected, and you watch the Handler_* counters, you can see how much work has been done.

]]> By: Mekin http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286511 Mekin Wed, 23 Apr 2008 13:25:51 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286511 Is it possible to estimate this for - inserts, updates & deletes - alter table queries Thanks, mekin Is it possible to estimate this for

- inserts, updates & deletes
- alter table queries

Thanks,
mekin

]]> By: Baron Schwartz http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286452 Baron Schwartz Wed, 23 Apr 2008 11:22:41 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286452 Hi Mannish, You should ignore the first set of outputs from the command. The second and subsequent outputs will contain incremental numbers per ten seconds. You can't see from the status variables how many connections have ever been created, but you can log in and execute SELECT CONNECTION_ID() for a rough idea. This value may wrap around to zero, though. Sometimes queries show up in the slow logs because of other things that are happening, such as backups or a lock by another query. Hi Mannish,

You should ignore the first set of outputs from the command. The second and subsequent outputs will contain incremental numbers per ten seconds.

You can’t see from the status variables how many connections have ever been created, but you can log in and execute SELECT CONNECTION_ID() for a rough idea. This value may wrap around to zero, though.

Sometimes queries show up in the slow logs because of other things that are happening, such as backups or a lock by another query.

]]> By: Manish Poddar http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/comment-page-1/#comment-286264 Manish Poddar Wed, 23 Apr 2008 05:14:14 +0000 http://www.mysqlperformanceblog.com/2008/04/22/how-to-estimate-query-completion-time-in-mysql/#comment-286264 Hi Baron, I was wondering if you were to take the value of Handler_read_rnd_next and divide it by number of total threads connected (ever?) will that give us a reasonable row reads per second? When I ran "mysqladmin extended -r -i 10 | grep Handler_read_rnd_next" I got a value of 1849842923. | Threads_connected | 9 | | Threads_created | 1042 | | Threads_running | 1 | | Uptime | 6180733 | so either I am getting too good row reads per second or I am misled by the data. What should I read into when the value of Handler_read_rnd_next is at 184M. We do have the slow query log enabled and we sometimes see the queries (that are properly indexed) show up in the slow logs. Any help in this matter will be greatly appreciated. Hi Baron,

I was wondering if you were to take the value of Handler_read_rnd_next and divide it by number of total threads connected (ever?) will that give us a reasonable row reads per second? When I ran “mysqladmin extended -r -i 10 | grep Handler_read_rnd_next” I got a value of 1849842923.

| Threads_connected | 9 |
| Threads_created | 1042 |
| Threads_running | 1 |
| Uptime | 6180733 |

so either I am getting too good row reads per second or I am misled by the data.

What should I read into when the value of Handler_read_rnd_next is at 184M. We do have the slow query log enabled and we sometimes see the queries (that are properly indexed) show up in the slow logs.

Any help in this matter will be greatly appreciated.

]]>