Summary:	ASTERISK-13439: [patch] describe idlecheck in res_odbc.conf.sample more clear
Reporter:	klaus3000 (klaus3000)	Labels:
Date Opened:	2009-01-23 01:20:53.000-0600	Date Closed:	2011-06-07 14:08:06
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Documentation
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) patch_idlecheck_res_odbc.conf.sample.txt
Description:	Hi! I did not understand correctly what this parameter does unless reading the code. I think adding this text will make things more clear. ****** ADDITIONAL INFORMATION ****** I think applying to 1.4 branch would be useful too.
Comments:	By: Mark Michelson (mmichelson) 2009-01-23 14:52:28.000-0600 This text seems like a good addition. I will commit it as is, except that I will correct the misspelled word "secondes" to be "seconds." Thanks for the documentation update! By: Digium Subversion (svnbot) 2009-01-23 14:55:14.000-0600 Repository: asterisk Revision: 170719 U   branches/1.4/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170719 | mmichelson | 2009-01-23 14:55:13 -0600 (Fri, 23 Jan 2009) | 8 lines Add notes to the idlecheck explanation in res_odbc.conf.sample (closes issue ASTERISK-13439) Reported by: klaus3000 Patches:      patch_idlecheck_res_odbc.conf.sample.txt uploaded by klaus3000 (license 65) ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170719 By: Digium Subversion (svnbot) 2009-01-23 14:55:40.000-0600 Repository: asterisk Revision: 170720 _U  trunk/ U   trunk/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170720 | mmichelson | 2009-01-23 14:55:40 -0600 (Fri, 23 Jan 2009) | 16 lines Merged revisions 170719 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r170719 | mmichelson | 2009-01-23 14:55:26 -0600 (Fri, 23 Jan 2009) | 8 lines Add notes to the idlecheck explanation in res_odbc.conf.sample (closes issue ASTERISK-13439) Reported by: klaus3000 Patches:      patch_idlecheck_res_odbc.conf.sample.txt uploaded by klaus3000 (license 65) ........ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170720 By: Digium Subversion (svnbot) 2009-01-23 14:56:05.000-0600 Repository: asterisk Revision: 170721 _U  branches/1.6.0/ U   branches/1.6.0/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170721 | mmichelson | 2009-01-23 14:56:05 -0600 (Fri, 23 Jan 2009) | 24 lines Merged revisions 170720 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................ r170720 | mmichelson | 2009-01-23 14:56:07 -0600 (Fri, 23 Jan 2009) | 16 lines Merged revisions 170719 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r170719 | mmichelson | 2009-01-23 14:55:26 -0600 (Fri, 23 Jan 2009) | 8 lines Add notes to the idlecheck explanation in res_odbc.conf.sample (closes issue ASTERISK-13439) Reported by: klaus3000 Patches:      patch_idlecheck_res_odbc.conf.sample.txt uploaded by klaus3000 (license 65) ........ ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170721 By: Digium Subversion (svnbot) 2009-01-23 14:57:14.000-0600 Repository: asterisk Revision: 170722 _U  branches/1.6.1/ U   branches/1.6.1/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170722 | mmichelson | 2009-01-23 14:57:13 -0600 (Fri, 23 Jan 2009) | 24 lines Merged revisions 170720 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................ r170720 | mmichelson | 2009-01-23 14:56:07 -0600 (Fri, 23 Jan 2009) | 16 lines Merged revisions 170719 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r170719 | mmichelson | 2009-01-23 14:55:26 -0600 (Fri, 23 Jan 2009) | 8 lines Add notes to the idlecheck explanation in res_odbc.conf.sample (closes issue ASTERISK-13439) Reported by: klaus3000 Patches:      patch_idlecheck_res_odbc.conf.sample.txt uploaded by klaus3000 (license 65) ........ ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170722 By: Tilghman Lesher (tilghman) 2009-01-23 15:00:20.000-0600 While this note is technically true, I fail to see why a user needs that level of information about the mechanics of reconnecting. By: Digium Subversion (svnbot) 2009-01-24 07:54:34.000-0600 Repository: asterisk Revision: 170836 U   branches/1.4/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170836 | tilghman | 2009-01-24 07:54:33 -0600 (Sat, 24 Jan 2009) | 2 lines Remove superfluous implementation note (closes issue ASTERISK-13439) ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170836 By: Digium Subversion (svnbot) 2009-01-24 07:55:24.000-0600 Repository: asterisk Revision: 170837 _U  trunk/ U   trunk/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170837 | tilghman | 2009-01-24 07:55:24 -0600 (Sat, 24 Jan 2009) | 9 lines Merged revisions 170836 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........  r170836 | tilghman | 2009-01-24 07:55:02 -0600 (Sat, 24 Jan 2009) | 2 lines    Remove superfluous implementation note (closes issue ASTERISK-13439) ........ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170837 By: Digium Subversion (svnbot) 2009-01-24 07:56:13.000-0600 Repository: asterisk Revision: 170838 _U  branches/1.6.0/ U   branches/1.6.0/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170838 | tilghman | 2009-01-24 07:56:12 -0600 (Sat, 24 Jan 2009) | 16 lines Merged revisions 170837 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................  r170837 | tilghman | 2009-01-24 07:55:53 -0600 (Sat, 24 Jan 2009) | 9 lines    Merged revisions 170836 via svnmerge from  https://origsvn.digium.com/svn/asterisk/branches/1.4    ........    r170836 | tilghman | 2009-01-24 07:55:02 -0600 (Sat, 24 Jan 2009) | 2 lines        Remove superfluous implementation note (closes issue ASTERISK-13439)  ........ ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170838 By: Digium Subversion (svnbot) 2009-01-24 07:56:51.000-0600 Repository: asterisk Revision: 170839 _U  branches/1.6.1/ U   branches/1.6.1/configs/res_odbc.conf.sample ------------------------------------------------------------------------ r170839 | tilghman | 2009-01-24 07:56:51 -0600 (Sat, 24 Jan 2009) | 16 lines Merged revisions 170837 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................  r170837 | tilghman | 2009-01-24 07:55:53 -0600 (Sat, 24 Jan 2009) | 9 lines    Merged revisions 170836 via svnmerge from  https://origsvn.digium.com/svn/asterisk/branches/1.4    ........    r170836 | tilghman | 2009-01-24 07:55:02 -0600 (Sat, 24 Jan 2009) | 2 lines        Remove superfluous implementation note (closes issue ASTERISK-13439)  ........ ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=170839 By: klaus3000 (klaus3000) 2009-01-25 08:11:20.000-0600 Honestly I do not understand why you reject documentation. I (as user) use odbc. As I had some connection problems I was looking for some parameters for tweaking. As I found the idlecheck parameter I read its documentation and thought that it can solve my problems - instead it made things worse. Only after studying the source code I really understood what this parameter is about. IMO it is very essential that the connection checking is not done in background, but instead only when a query happens. You are talking that a user does not need this kind of information - obviously this is wrong. I am a user an I needed this information which was not available. Asterisk still misses lots of documentation - especially all the unixODBC stuff. And now you get a detailed documentation and reject it. I really do not understand this behavior. By: Tilghman Lesher (tilghman) 2009-01-25 12:04:30.000-0600 Perhaps you could describe the problem you are having, and we could suggest an approach to take.  Getting angry about inappropriate documentation does not solve any problems. By: klaus3000 (klaus3000) 2009-01-26 03:19:23.000-0600 My setup is: Asterisk 1.4.22/23 and unixODBC 2.2.11-16 and libmyodbc 3.51.15r409-4 and mysql server 5.0.51a-21. My problems were, that very often a query failed, the connection was closed, a new connection was established and then the query succeeded. And the problem was that the reconnecting always took 5 seconds (I guess a timeout somewhere in libodbc). So I began tweaking connection settings. The idlecheck solved the 5 seconds timeout issue, but introduced a IMO not nice behavior (disconnect/connect). So I tried other things. Now I used a TCP connection to the MySQL server and updated to Asterisk 1.4.23 (which has some ODBC changes) and until now the DB connection works fine. So, my problems are solved now, nevertheless I thought extending the documentation is a good idea to prevent other users to missnderstand the meaning of the idlecheck parameter (just as I did). Thus, back to the topic - I do not understand why you reject this IMO useful documentation. users which do not have connection problem will not use the parameter any. Users which run into CD issues will try to find some parameters for tuning, and these users for sure are interested what the idlecheck parameter is about in detail. By: Tilghman Lesher (tilghman) 2009-01-26 15:42:02.000-0600 So, in fact, the idlecheck parameter was introduced specifically for your issue, and it did solve it.  So I'm not sure what you're complaining about.  The only reason that you seem to have a problem with it is that you don't "like" the solution, for whatever reason.  But the solution works.  So I don't understand why the option confused you, as you found it easily, and it worked to solve your problem. "I don't like it" is not a valid reason to add documentation; it's merely your opinion. By: Tilghman Lesher (tilghman) 2009-01-26 15:45:37.000-0600 Additionally, adding information on the way things work 1) does not help most users, 2) gets in the way of most users figuring out what they need to configure, and 3) sets a bad precedent on where developer documentation should go (NOT in the configuration files). By: klaus3000 (klaus3000) 2009-01-27 12:46:36.000-0600 Just be sure that both of us know what we are talking about: +; Note: The reconnection is not done in the background, but is triggered by +; a new query. Thus, if you set idlecheck => X, every time you perform a +; query res_odbc will check how long these connection was idle, and if it was +; idle for more than X secondes it will close the current connection and +; establish a new one even if the current connection is still working. This is not a developer documentation but a verbose user documentation. > Additionally, adding information on the way things work 1) does > not help most users But it help some of the user. But we do not care about "some" ??? > 2) gets in the way of most users figuring out what they need to configure If users should not configure it - remove it. If you have a parameter then document it that all users understand it. > So, in fact, the idlecheck parameter was introduced specifically > for your issue, and it did solve it. I still do not know what my issue was exactly - but it was solved by updating to 1.4.23. idlecheck did not solved the issue - it just was an ugly workaround. Nevertheless I am tired of contributing and arguing for documentation. If you think that you are more user than me and understand user's concern better then close the bug again - I do not have that much idle time.