*** ii_log has joined #ingres | 00:14 | |
*** ii_log` has joined #ingres | 01:22 | |
*** ii_log has quit IRC | 01:35 | |
*** rossand has joined #ingres | 02:56 | |
*** ChanServ sets mode: +o rossand | 02:56 | |
*** rossand has quit IRC | 03:04 | |
*** perftuning has quit IRC | 04:22 | |
*** toumi01 has quit IRC | 04:39 | |
*** Alex|off is now known as Alex| | 07:10 | |
*** Gerhard has joined #ingres | 07:47 | |
Gerhard | HI all | 07:50 |
---|---|---|
grantc | hi Gerhard | 07:55 |
grantc | @nick ii_log | 07:55 |
*** ii_log` is now known as ii_log | 07:55 | |
Gerhard | Any experiences with the "dotnet data provider" ingres.client.dll? | 08:01 |
grantc | not really, i have played with it a couple of times | 08:02 |
Gerhard | I like the idea of having only a single DLL on the client instead of a full-blown Ingres Client installation | 08:05 |
grantc | its a bit more involved than a single dll but yes I understand what you are saying, like with iijdbc.jar | 08:06 |
grantc | i wish we could do something similar for ODBC/PHP/Python/Ruby | 08:10 |
mc349 | We are starting to use it extensively and are very impressed. It is just a matter of dropping one DLL into the GAC and you are good to go. (As long as you've got the .net framework stuff already installed.) | 08:14 |
grantc | nice - IronPython seems to work with it as well | 08:15 |
mc349 | It's faster than IngresNet / ODBC as it cuts out a few layers. Unfortunately the iigcd process on the back end isn't load balancable like iigcc is. But I believe I've seen that feature on a planned enhancment list somwhere. | 08:16 |
mc349 | Another bonus is having the source code to look at and even step through. I don't think that's available for the client side ODBC driver yet? | 08:18 |
grantc | mc349, its in the source tree for ingres under common/odbc | 08:19 |
mc349 | Oh good, I'll have a look at that, when I get some time ;-) | 08:20 |
grantc | although if you want to debug the C the you should build ODBC/ODBC CLI/OpenAPI and perhaps parts of the CL in debug :) | 08:20 |
grantc | ;) | 08:21 |
grantc | btw the project you mentioned is http://community.ingres.com/wiki/Multiple_DAS_Support_Home | 08:21 |
mc349 | We've been using the ODBC driver successfully for at least 8 years, I doubt there are any bugs left for us to find by now. | 08:21 |
grantc | :) | 08:22 |
Alex| | can we make a success story out of that? :) | 08:23 |
mc349 | The one thing that I wish they both would support would be the Ingres copy from / copy into statements. For that we have to go back to IngresNet based stuff. | 08:24 |
Gerhard | grantc: you mentioned that it would be nice to have an ODBC driver that could directly connect to DAS, I'd like to have this also... | 08:38 |
pboro | me too | 08:40 |
Gerhard | ...would it be possible to write an ODBC wrapper around ingres.client.dll so that older apps (like Crystal 8.5) - that can only speak ODBC - could use ingres.client.dll? | 08:41 |
grantc | there are several issues, the JDBC/.NET drivers depend on features in the host language for date/time and charsets | 08:41 |
grantc | Gerhard, if you could write an C wrapper, yes | 08:42 |
grantc | a colleague started to look at writing a pure python driver that interfaced with the DAS server, not sure how far he got | 08:43 |
Gerhard | It's quite frustrating that "ODBC-against-Ingres" is such an installation hassle compared to "JDBC-against-Ingres" which can even run from UNC paths, USB sticks etc. | 08:49 |
pboro | Gerhard, btw have you tried JDBC with Crystal and Ingres? | 08:50 |
pboro | Just wondering if I should switch from ODBC to JDBC with Crystal... | 08:50 |
Gerhard | No, I have never tested the Crystal + JDBC combination... | 08:51 |
pboro | okay | 08:52 |
Gerhard | our own apps are OpenRoad with usage of Crystal 8.5 with ODBC (quite a lot of stuff to install)... | 08:52 |
Gerhard | ...but we have a third-party app running JDBC against Ingres, which is nice as far as installation is concerned (does not need an Ingres client setup) | 08:53 |
pboro | yup | 08:53 |
Gerhard | I think, some third-party ODBC vendors like OpenLink made Ingres-ODBC-drivers in the past that did not need a full-blown client setup... | 08:55 |
Gerhard | ...but on the other hand needed some additional "server software" | 08:55 |
pboro | oh, I didn't know that | 08:55 |
pboro | doh :) | 08:55 |
pboro | so no big bonus with that either | 08:56 |
Gerhard | you're right, but why not provide an Ingres-ODBC-driver that connects to DAS? Or asked in another way: why isn't it possible for Ingres Corp to provide a lightweight ODBC driver... | 09:00 |
Gerhard | ...when they already provide a lightweight JDBC driver and lightweight Dotnet Data provider? | 09:01 |
Gerhard | Just because ODBC is "old"? | 09:01 |
grantc | openlink had/have a setup similar to the DAS | 09:02 |
grantc | not sure it is a question of age... | 09:03 |
grantc | if you really want it, and I know you both do, raise an issue... | 09:04 |
grantc | Alex|, check out http://www.bombstrike.org/2009/02/bringing-javascript-to-the-server/ | 09:04 |
grantc | not sure if that might be of any use | 09:05 |
pboro | we use javascript on server side for several things | 09:07 |
pboro | not for SQL tho... | 09:07 |
grantc | it's not a PECL module yet but was brought up on the mailing list this morning | 09:07 |
pboro | but simply for system users to be able to customize some mathematical calculations, for example we do calcs for students' grading and every teacher may decide to use his/her own grading calculations :) | 09:08 |
pboro | there are a number of templates for common cases and if none of them suits, the user can write the calcs on his/her own... it's actually quite cool :) | 09:08 |
grantc | update grades set grade = grade + 10 .... | 09:08 |
pboro | :D | 09:08 |
Gerhard | Good idea, I'll raise an issue with tech support. | 09:09 |
grantc | or bobby drop tables :) | 09:10 |
pboro | grantc, JS has no external context, you can import some values into the javascript scope and then after JS execution, export them from it | 09:10 |
pboro | so it's perfectly safe | 09:10 |
pboro | (as far as there's no holes in the JS engine itself) | 09:10 |
grantc | :) | 09:11 |
grantc | has anyone here read code complete? | 09:22 |
pboro | Gerhard, I raised an issue about the odbc driver, I wonder what will the answer be :) | 10:06 |
grantc | if you want something from Ingres Corp it's the best way to get any action. For certain things you can poke a dev directly but ultimately we do what we are told by the higher-ups | 10:08 |
pboro | yup | 10:09 |
pboro | I guess that's the way it should be :) | 10:09 |
grantc | indeed - as a former colleague once said, if it's not written down it does not exist :) | 10:10 |
pboro | I hope he's not former because of saying that ;D | 10:11 |
pboro | I guess ITIL is used pretty widely these days, even in software development | 10:12 |
pboro | ie. http://en.wikipedia.org/wiki/ITIL | 10:12 |
grantc | no, in fact he was former when he said it :) | 10:13 |
pboro | oh okay :D | 10:13 |
pboro | nice for him | 10:13 |
grantc | http://croker.net/~ii_log/irc/logs/%23ingres/ is now being updated every two hours... | 10:14 |
Gerhard | pboro: thanks, I also opened an issue at service desk. My issue number is 134502. | 11:07 |
Gerhard | Maybe we could "cross link" our issues to emphasize the importance. | 11:07 |
pboro | I don't know my issue number yet 'cos we have to do all our service desk stuff through a local support company, but I'll keep this in mind | 11:08 |
pboro | Gerhard, 134508 | 11:20 |
pboro | Magnus from Sweden's Tieto had never heard about this kind of enhancement request, which amazes me a bit. Or not :) | 11:27 |
pboro | I probably look into the JDBC stuff with Crystal if that would work instead of ODBC stuff | 11:27 |
grantc | pboro, i guess it depends on what end of the stick you work ... | 11:28 |
pboro | hehe, yeah :) | 11:29 |
grantc | vi vs emacs, devs vs dbas... | 11:29 |
pboro | I started to wonder whether #ingres should have a live meeting some day! :) | 11:30 |
pboro | unfortunately there are so few of us so I guess it wouldn't make much sense | 11:30 |
grantc | i guess it depends on what you want to talk about? | 11:31 |
pboro | yeah I was thinking more of social stuff than some real conversations about some specific topics :D | 11:33 |
grantc | ingres drinking game ... for every sql statement mentioned you have to down a mouthful? | 11:34 |
pboro | hehe | 11:35 |
pboro | I guess the people on the chan are so few and so scattered that having a social get-to-know or similar meeting is not possible | 11:35 |
grantc | perhaps | 11:36 |
*** troal01 has left #ingres | 12:40 | |
*** troal01 has joined #ingres | 12:45 | |
*** perftuning has joined #ingres | 13:03 | |
elPiola | grantc, re Code Complete, I have, eons ago | 13:16 |
grantc | ok - i keep hearing about it on the stack overflow podcast | 13:17 |
grantc | was wondering if it was any good? | 13:17 |
elPiola | from what I remember it was thick, so I'm not sure if I finished it, and I sold it or gave it away | 13:18 |
grantc | :) | 13:18 |
elPiola | I have kept Writing Solid Code, but I haven't looked at it for quite some time | 13:19 |
Gerhard | What is the difference between "wintcp" and "tcp_ip" when adding a Vnode in VisualDBA? | 13:34 |
grantc | the latter uses the new Winsock2 protocol driver | 13:35 |
Gerhard | Thanks. Is it faster than the older protocol? | 13:40 |
*** rossand has joined #ingres | 13:52 | |
*** ChanServ sets mode: +o rossand | 13:52 | |
*** Gerhard has quit IRC | 13:59 | |
mc349 | can "customers" who are not paying for support raise a service request / enhancement request? | 14:22 |
Alex| | at http://bugs.ingres.com | 14:23 |
Alex| | the community tracking system | 14:23 |
troal01 | BTW grantc, this lets me exceed the page limitation: modify quote to reconstruct with partition = (hash on id 4 partitions (part1, part2, part3, part4)), location = (data_1, data_2, data_3, data_4)\g | 14:26 |
troal01 | So I guess it's confirmed just reorganizing to other locations isn't sufficient to grow the table | 14:26 |
grantc | ok - glad you worked it out | 14:26 |
grantc | now can you do a write up for the wiki? | 14:26 |
troal01 | Hehehe, yeah sure, where do you want me to put it? | 14:27 |
grantc | with tid > 2^23 | 14:27 |
grantc | do the write up and I will have a think about it | 14:28 |
grantc | i guess under Articles.. | 14:28 |
troal01 | So there's not a page for it currently? | 14:29 |
grantc | you could write it as a how to... or what to do when ... | 14:30 |
troal01 | Under tech tips is probably a good place | 14:31 |
troal01 | Do we have code tags for our wiki? | 14:45 |
troal01 | I guess <pre> tags will have to do | 14:50 |
troal01 | Would be nice to have the <source> extension though | 14:50 |
troal01 | http://community.ingres.com/wiki/TID_Limitation | 14:53 |
*** DarylM has joined #ingres | 14:55 | |
troal01 | Hey DarylM, mind taking a look? http://community.ingres.com/wiki/TID_Limitation | 14:56 |
DarylM | k | 14:56 |
DarylM | I don't have the reference handy but I believe there is a fairly exact formula on rows/page in the dba guide that would be a nice reference there. | 14:58 |
troal01 | Let me do a quick search | 14:59 |
DarylM | Also, I don't think the locations are required? If so, it might be misleading because adding locations is a much bigger pain for a dba than partitioning a table. | 14:59 |
DarylM | (In fact, we are rapidly moving away from database locations. I consider them a leftover from pre-Raid/pre-SAN days) | 15:00 |
troal01 | Good question I didn't try without locations... But I guess they might not be necessary | 15:01 |
troal01 | Well I kinda like being able to locate my data on a separate HDD from my OS :) | 15:10 |
troal01 | Is this what you were talking about DarylM? http://docs.ingres.com/Ingres/9.2/Database%20Administrator%20Guide/calculatingdiskspace.htm#o1780 | 15:13 |
DarylM | Moving your data to a different file system from other stuff can be a good idea. Using multiple data locations isn't really necessary is the vast majority of cases anymore | 15:16 |
DarylM | There are of course exceptions | 15:16 |
DarylM | ;) | 15:16 |
DarylM | Ya | 15:16 |
troal01 | Great I'll link it in | 15:17 |
troal01 | Thanks | 15:17 |
DarylM | Right. The actual formula was embedded in the iitables system catalog. | 15:17 |
grantc | troal01, i am going to rename the page to "Working with large tables" | 15:33 |
grantc | assuming you have no objections :) | 15:33 |
troal01 | Not at all, I'm terribly at naming :) | 15:40 |
*** Alex| is now known as Alex|off | 15:42 | |
grantc | :) | 15:42 |
grantc | done | 15:42 |
grantc | http://community.ingres.com/wiki/Working_with_large_tables | 15:42 |
DarylM | Succinct and to the point. I like it. | 15:44 |
troal01 | Thanks grantc, good reference :) | 15:48 |
grantc | btw the IRC logs of this channel are now at http://irc.planetingres.org | 15:54 |
grantc | updated every 30 mins or so | 15:54 |
DarylM | ZOMG we are enshrined forever........ | 15:55 |
*** grantc is now known as grantc_offline | 15:57 | |
troal01 | LOL | 15:57 |
*** grantc_offline is now known as grantc | 16:00 | |
grantc | troal01, are you going to extend that wiki page to include the steps you need to create the locations and extend the db? | 16:06 |
troal01 | No because according to DarylM the extra locations aren't necessary | 16:07 |
troal01 | I never tested but I trust his memory! :) | 16:07 |
DarylM | foolish mortal | 16:07 |
troal01 | Spare me oh awesome guru of Ingres | 16:08 |
* DarylM is sitting here Yoga style in his chair but not because he is a guru, but because of the level of BS on the floor | 16:09 | |
DarylM | ;) | 16:09 |
troal01 | Hehehe | 16:09 |
mc349 | is there a maximum number of partitions for a table? Perhaps this limit (if any) should be mentioned on the wiki page, giving the maximum theoretical rows that Ingres can handle in a single table. Not that I like growing anything above a few tens of millions! | 16:18 |
mc349 | (of rows) | 16:19 |
troal01 | That's a good question mc349, DarylM any idea? | 16:21 |
DarylM | That is a good question. The SQL reference manual does have the following note" Note: If tables are defined with thousands of partitions, Ingres should be | 16:24 |
DarylM | properly configured." | 16:24 |
mc349 | The DBA should be properly reconfigured in my opinion! | 16:25 |
DarylM | :D | 16:25 |
troal01 | LOL | 16:27 |
*** toumi01 has joined #ingres | 16:41 | |
*** Alex|off is now known as Alex| | 17:00 | |
*** Alex| is now known as Alex|off | 21:00 | |
mc349 | Hi Roy! | 21:40 |
*** ukjay has joined #ingres | 21:40 | |
*** ukjay has left #ingres | 21:41 | |
*** DarylM has quit IRC | 23:35 | |
*** grantc is now known as grantc_offline | 23:55 | |
*** elPiola has left #ingres | 23:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!