*** tvstebut has joined #ingres | 00:23 | |
tvstebut | Hello | 00:23 |
---|---|---|
pboro | hi tvstebut | 00:24 |
tvstebut | I want to try ingres but I'd like to know first how Ingres DB Cmmunity Edition is different from MySQL ???? | 00:24 |
tvstebut | And I can't find this information on the web | 00:25 |
tvstebut | And my second question is about difference between Ingres Community edition and the Enterprise Edition ....? | 00:25 |
tvstebut | Hi pboro | 00:25 |
pboro | in what sense? they both support SQL language, but Ingres has far more features | 00:25 |
tvstebut | I should have asked the questions at Talend SUmmit last week...there was a Ingres stand | 00:26 |
tvstebut | In the feature sense actually | 00:26 |
pboro | there's no commercial support for the community edition, the source code is released freely under gpl, community edition does not receive patches or the same testing as the commercial releases etc. | 00:28 |
tvstebut | ok, but what about features in details ? | 00:29 |
pboro | if you use MySQL with myisam and you want real relational dmbs, then ingres wins 100-0 :) since MySQL doesn't provide referential consistency etc. when using myisam | 00:29 |
pboro | with innodb things are much more even, but still, innodb lacks some features provided by ingres... and so on, it's hard to give a precise list of differences since there are so many small things | 00:30 |
pboro | you are familiar with MySQL's storage engines, aren't you? | 00:31 |
tvstebut | Yes I am | 00:32 |
tvstebut | as Oracle also | 00:32 |
pboro | yeah | 00:32 |
*** Alex| has joined #ingres | 00:32 | |
*** ChanServ sets mode: +o Alex| | 00:32 | |
tvstebut | okay | 00:32 |
pboro | what specific features are you interested in? | 00:32 |
tvstebut | That would be good to find a detailed comparaison matrix somewhere :p | 00:33 |
pboro | I'm sure that overall they both do their work, both support SQL and tables and so on :) | 00:33 |
tvstebut | :-) | 00:33 |
tvstebut | hmmm | 00:33 |
tvstebut | okay I want also to be able to do some partitionning etc... explain plans....sort of things | 00:34 |
tvstebut | And....clustering....failover....etc | 00:34 |
tvstebut | Thats why I'm also looking forward to know if Ingres Community Edition lacks or not some of the Enterprise edition features...? | 00:35 |
pboro | Ingres does support table partitioning (http://docs.ingres.com/Ingres/9.3/SQL%20Reference%20Guide/modify.htm), and it's possible to view the query execution plan for a query | 00:36 |
pboro | Emma McGrattan from Ingres wrote in 2008 that "...we haven't held anything back from the community version of Ingres - if you want replication, distributed transaction process, high availability, hot backups, the latest set of features etc. etc. you'll find them in both the community and the certified products. The code for both versions is identical - where they differ is in the level of quality assurance, performance and stress testing we put them through" | 00:36 |
pboro | and that's still the case as far as I know | 00:37 |
tvstebut | Ok great | 00:37 |
tvstebut | but no partitioning...whereas mySQL does :-/ | 00:37 |
tvstebut | oh you wrote does | 00:38 |
tvstebut | my fault | 00:38 |
pboro | yeah partitioning is there | 00:38 |
tvstebut | I red too fast ;) | 00:38 |
tvstebut | goood | 00:38 |
tvstebut | definitly | 00:38 |
tvstebut | Know I think I have a strong state of mind, well enough to proudly install and test this database in my project :) | 00:39 |
tvstebut | I use the linux quick start | 00:39 |
pboro | by browsing the manual you can get an idea of what is supported and what's not... i don't expect you to find any big empty spots :) | 00:39 |
tvstebut | Thank you pboro :-) that rocks | 00:40 |
pboro | I've been using Ingres for web applications for some years now and it has worked quite nicely, some bugs have been encountered but they have been fixed by the support | 00:40 |
pboro | overall the quality of ingres has been better and better all the time, in 2005 there was some quality problems in terms of number of bugs etc. but the testing and qa stuff in ingres has been getting better | 00:41 |
pboro | I guess one challenging area is replication if you just want to use the community editions, since ingres does not contain modern replication features integrated in the product but as a separate product from 3rd party | 00:42 |
pboro | http://www.hvr-software.com/news/news_20091030184226 | 00:43 |
pboro | and that's not free, it's a commercial-only product | 00:44 |
pboro | so in that sense ingres is lacking a feature that's provided by mysql | 00:45 |
tvstebut | pboro, can't I just use Rsync on datafile ? | 00:47 |
tvstebut | Is that doable | 00:51 |
pboro | aaa sure | 01:05 |
pboro | you can dump the database into files | 01:05 |
pboro | but I meant like realtime replication over tcp/ip or similar | 01:06 |
tvstebut | hmmm...replication for what....failover ? | 01:06 |
pboro | yeah | 01:06 |
pboro | like hot-standby or similar | 01:06 |
tvstebut | aah so HVR is the only solution ? | 01:06 |
tvstebut | hot-standby yeah | 01:07 |
pboro | yup | 01:07 |
tvstebut | That is bothering right now | 01:07 |
tvstebut | :-/ | 01:07 |
tvstebut | There could be free solutions but no | 01:08 |
pboro | yeah I've been thinking whether to write a one by myself, but I have no time do such a big projects | 01:09 |
tvstebut | with Rsync there would be a little down time I guess | 01:09 |
tvstebut | But its doable, I mean a crontab that replicates the whole dirs every 5 minutes or somethong tunable | 01:10 |
tvstebut | what do you think about that ? | 01:10 |
tvstebut | How does it work whith MySQL ? | 01:10 |
pboro | it's realtime, you designate one db as a master and one as a slave and then it just... works :) | 01:11 |
tvstebut | Hey, now I don't know what to choose doh | 01:12 |
* tvstebut 's upset | 01:12 | |
pboro | http://dev.mysql.com/doc/refman/5.5/en/replication.html | 01:12 |
pboro | there's the replication part from mysql manual | 01:12 |
tvstebut | okay | 01:13 |
tvstebut | Well I think I will choose MySQL then | 01:13 |
tvstebut | That's serious, despite my wanting give try to Ingres :-/ | 01:13 |
pboro | yeah that's something that disturbs me too a bit | 01:14 |
*** Alex| has quit IRC | 01:18 | |
tvstebut | What about IngresReplicator | 01:18 |
tvstebut | ? | 01:18 |
*** Alex| has joined #ingres | 01:19 | |
*** ChanServ sets mode: +o Alex| | 01:19 | |
pboro | it's trigger-based, so it's pretty heavy and outdated solution... | 01:19 |
pboro | most replication solutions nowadays are based on transaction logging, the log data is transferred over the wire and applied to other dbs | 01:20 |
tvstebut | :/ | 01:21 |
*** Alex| has quit IRC | 01:22 | |
*** Alex| has joined #ingres | 01:22 | |
*** ChanServ sets mode: +o Alex| | 01:22 | |
*** Alex| has quit IRC | 01:33 | |
*** Alex| has joined #ingres | 01:35 | |
*** ChanServ sets mode: +o Alex| | 01:35 | |
*** Alex| has quit IRC | 01:42 | |
*** Alex| has joined #ingres | 01:42 | |
*** ChanServ sets mode: +o Alex| | 01:42 | |
*** Alex| has quit IRC | 01:46 | |
*** Alex| has joined #ingres | 01:50 | |
*** ChanServ sets mode: +o Alex| | 01:50 | |
*** Alex| has quit IRC | 01:55 | |
*** Alex| has joined #ingres | 01:56 | |
*** ChanServ sets mode: +o Alex| | 01:56 | |
*** Alex| has quit IRC | 02:14 | |
tvstebut | pboro, hi again | 02:18 |
tvstebut | When googling about replication methods, it doesn't appear that log-based replication is better | 02:18 |
tvstebut | not roughly | 02:19 |
tvstebut | pboro, you are maybe a little too "perfectionist" | 02:22 |
tvstebut | :) | 02:24 |
pboro | it depends | 02:28 |
pboro | trigger-based solution is a bit of ad-hoc if I'm asked, since all the necessary data etc. could easily be collected from a transaction log | 02:29 |
tvstebut | So you'd recommend log-based model for this reason | 02:30 |
pboro | not only for that, if you read from ingres manual about ingres replicator, you'll find out that it does not pick all changes to the db | 02:30 |
pboro | for example if you change schema... | 02:30 |
tvstebut | oh | 02:30 |
tvstebut | thats a point | 02:31 |
*** Dejan has joined #ingres | 02:35 | |
Dejan | good day | 02:36 |
tvstebut | Yes right | 02:39 |
tvstebut | now pboro give me reasons to chose infres agains lack of replication robusness ;) | 02:40 |
tvstebut | Hi Dejan | 02:40 |
Dejan | tvstebut, there is a product called HVR | 02:43 |
tvstebut | Dejan, is it free ? | 02:43 |
Dejan | which is used by big Ingres customers | 02:43 |
Dejan | no, HVR is a commercial product | 02:43 |
tvstebut | hmmm, ok, but I'm interrested only by community edition | 02:43 |
pboro | tvstebut, I'm not going to take any sides, I'm just a customer :) | 02:44 |
Dejan | my noobish opinion is - replication can actually be implemented very easily | 02:44 |
Dejan | a pretty good one | 02:44 |
*** tvstebut_ has joined #ingres | 02:44 | |
Dejan | tvstebut, did you read what i wrote? | 02:45 |
tvstebut_ | that its a commercial product ? | 02:45 |
Dejan | <Dejan> my noobish opinion is - replication can actually be implemented very easily | 02:46 |
Dejan | <Dejan> a pretty good one | 02:46 |
tvstebut_ | okays thanks | 02:46 |
tvstebut_ | so how can it be ? (without HVR) | 02:46 |
Dejan | you can periodically export transactions into a file | 02:46 |
Dejan | copy it over network to a replica-server | 02:47 |
Dejan | and execute them there too | 02:47 |
Dejan | or write a tool which will do this automatically | 02:47 |
Dejan | or, ultimately, patch Ingres | 02:47 |
Dejan | :D | 02:47 |
*** tvstebut has quit IRC | 02:48 | |
*** tvstebut has joined #ingres | 02:49 | |
*** tvstebut_ has quit IRC | 02:51 | |
*** tvstebut_ has joined #ingres | 02:52 | |
tvstebut_ | Adding to this a soft load-balancer :) .... | 02:52 |
*** tvstebut has quit IRC | 02:53 | |
*** tvstebut_ has quit IRC | 02:58 | |
*** tvstebut has joined #ingres | 02:59 | |
*** tvstebut has quit IRC | 03:17 | |
*** tvstebut has joined #ingres | 03:18 | |
*** tvstebut has quit IRC | 03:22 | |
*** Alex| has joined #ingres | 05:32 | |
*** ChanServ sets mode: +o Alex| | 05:32 | |
*** rossand has joined #ingres | 05:58 | |
*** ChanServ sets mode: +o rossand | 05:58 | |
*** Alex| has quit IRC | 06:08 | |
FrankW | Is it still true that building ingres with gcc 4+ (4.4.1 in my case) is a problem? | 08:10 |
pboro | at least I'm building it with GCC 4 and it seems to work ok | 08:11 |
FrankW | http://community.ingres.com/wiki/Building_Ingres_On_Linux appears to be out of date. | 08:11 |
pboro | my automated builds: http://www.fixel.org/ingres/ | 08:11 |
FrankW | pboro: good enough for me. | 08:11 |
FrankW | I mean with regard to gcc4. I can't actually use your binary builds since I need the geospatial branch (and to build changes myself). | 08:12 |
FrankW | During my build I find this in my log file: | 08:13 |
FrankW | Link /home/warmerda/ingres/server/build/utility/unimapcompile | 08:13 |
FrankW | /usr/bin/ld: skipping incompatible /usr/local/32bit/lib/libgeos_c.so when searching for -lgeos_c | 08:13 |
FrankW | /home/warmerda/ingres/server/src/common/adf/adl/UNIMapCompile.o: In function `main': | 08:13 |
FrankW | UNIMapCompile.cpp:(.text+0xe2): undefined reference to `xercesc_2_7::XMLUni::fgXercescDefaultLocale' | 08:13 |
FrankW | How do I find out what the "Link" command is actually doing to link? | 08:13 |
pboro | hmmh :/ | 08:13 |
*** Alex| has joined #ingres | 08:15 | |
*** ChanServ sets mode: +o Alex| | 08:15 | |
pboro | FrankW, you do have xerces installed? | 08:18 |
FrankW | I have a system xerces3 in /usr and I built xerces 2.7 from source and installed in /usr/local (but 64bit only I imagine). | 08:19 |
FrankW | Do you do your 64bit builds in hybrid mode with 32bit and 64bit binaries? | 08:22 |
FrankW | actually, from what you wrote in your readme, you must. | 08:23 |
pboro | I do hybrid | 08:23 |
pboro | I can try building geos-tree if you wish | 08:23 |
pboro | with the same environment, dunno if it should work tho' :) | 08:23 |
FrankW | no, I'm not sure that would help me. | 08:23 |
FrankW | Does ingres really specifically require xerces 2.7? | 08:24 |
pboro | I read somewhere that it's not requirement anymore, but haven't tried it myself | 08:24 |
FrankW | Is there a way of seeing the actual link commands used? | 08:24 |
FrankW | I find it so frustrating that the actual commands used are hidden in the logs. | 08:25 |
FrankW | And it is so hard to find where link directives and such come from. | 08:25 |
pboro | there's some switches for jam to make it more verbose, but can't recall them right now... Alex might know | 08:25 |
FrankW | I've blown away my build directory and I'm doing "runbuild.sh" again. | 08:25 |
pboro | I'm using runbuild.sh -a | 08:26 |
pboro | what was -a for... hm hm | 08:26 |
pboro | ah i'm using buildrel, not runbuild.sh | 08:26 |
FrankW | I believe -a forces a rebuild. | 08:26 |
pboro | ok | 08:26 |
FrankW | Christ, this is driving me nuts. | 08:30 |
FrankW | I think I'm going to setup a 32bit VM and work in that. | 08:30 |
pboro | hehe :) | 08:30 |
FrankW | I will gently suggest that the build challenges of ingres are a significant drag on community developer interest. | 08:31 |
pboro | I agree completely :) | 08:31 |
pboro | there should at least be some kind of test system or similar which would check whether the build environment is ok | 08:32 |
pboro | ie. does it have all the necessary libs and does linking work ok etc. | 08:32 |
pboro | since for example xerces is repeating problem for all :/ | 08:33 |
FrankW | What is amazing is that ingres has so few external dependencies, but even those few are still such problems. | 08:33 |
pboro | yeah :D | 08:33 |
FrankW | It doesn't bode well for the geospatial branch merging into main, and adding a couple of geospatial dependencies (geos, proj.4). | 08:34 |
FrankW | Do you think that 32bit Fedora 10 would be a pretty easy dev platform? | 08:35 |
pboro | Hmm well for anything else sure but for Ingres... dunno, haven't tried it | 08:36 |
FrankW | any suggestion for an easy target environment? | 08:36 |
pboro | i've been using ubuntu, since most of the docs on the wiki is about it :) | 08:38 |
FrankW | will do- grabbing 8.10 server. | 08:41 |
*** Alex| has quit IRC | 10:35 | |
*** Alex| has joined #ingres | 10:36 | |
*** ChanServ sets mode: +o Alex| | 10:36 | |
*** Dejan has quit IRC | 10:45 | |
*** gerhard has joined #ingres | 12:46 | |
*** tvstebut has joined #ingres | 13:28 | |
*** Alex| has quit IRC | 13:33 | |
*** tvstebut has quit IRC | 13:40 | |
*** gerhard has quit IRC | 13:52 | |
*** epoitras has joined #ingres | 19:08 | |
*** cytrinox_ has joined #ingres | 19:21 | |
*** cytrinox has quit IRC | 19:24 | |
*** cytrinox_ is now known as cytrinox | 19:24 | |
*** epoitras has quit IRC | 19:47 | |
*** DarylM has joined #ingres | 22:17 | |
*** Alex| has joined #ingres | 22:40 | |
*** ChanServ sets mode: +o Alex| | 22:40 | |
*** grantc has joined #ingres | 22:47 | |
*** Alex| has quit IRC | 23:13 | |
*** grantc is now known as grantc_away | 23:18 | |
* grantc_away is away: please leave a message after the beep | 23:18 | |
*** Mud has joined #ingres | 23:45 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!