*** Dejan[off] is now known as Dejan | 00:55 | |
Dejan | morning | 00:55 |
---|---|---|
Dejan | this should be deleted from our forum: http://community.ingres.com/forum/migration-forum/11844-ingres-migration-tool-set.html#post30641 | 01:20 |
Dejan | it is totally irrellevant | 01:20 |
*** gerhard has joined #ingres | 01:31 | |
*** PaulM05 has joined #ingres | 01:32 | |
PaulM05 | good morning all | 01:36 |
Dejan | hi PaulM05 & gerhard | 01:55 |
Mud | yo dudesss :+ | 02:06 |
PaulM05 | hi Mud | 02:06 |
Dejan | hi Mud | 02:08 |
Dejan | :D | 02:08 |
Mud | hi ^^ | 02:10 |
gerhard | Hi all | 02:24 |
*** gerhard has quit IRC | 02:47 | |
Dejan | if i want to preserve all other parameters of a table, except page_size, do i just RECONSTRUCT it? | 02:49 |
PaulM05 | I believe so | 02:49 |
Dejan | ie: MODIFY tblname TO RECONSTRUCT WITH PAGE_SIZE=8192; | 02:49 |
Dejan | it preserves fillfactor etc ? | 02:50 |
PaulM05 | that's what it should do yes | 02:50 |
Dejan | ty | 02:50 |
*** withdefault has joined #ingres | 03:57 | |
*** ChanServ sets mode: +o withdefault | 03:57 | |
*** grantc has joined #ingres | 03:59 | |
grantc | morning all | 04:00 |
*** withdefault sets mode: -o withdefault | 04:01 | |
PaulM05 | hi grant | 04:01 |
withdefault | morning | 04:01 |
PaulM05 | morning withdefault | 04:01 |
withdefault | morning PaulM05 | 04:02 |
*** atrofast has joined #ingres | 04:14 | |
Dejan | morning | 04:17 |
Vroomfondle | hello | 04:18 |
grantc | Vroomfondle, have you or your sysadmin had a chance to look at the last change I made? | 04:19 |
Vroomfondle | Not yet | 04:25 |
Vroomfondle | I'm busy this week and haven't heard from him | 04:25 |
Vroomfondle | if he hasn't replied next week I can probably take a crack at it myself | 04:25 |
grantc | no problem - i'll put the issue on hold pending any feedback | 04:28 |
*** rossand has joined #ingres | 04:36 | |
*** ChanServ sets mode: +o rossand | 04:36 | |
*** Mud has quit IRC | 05:02 | |
*** PaulM05 is now known as PaulM05_lunch | 05:02 | |
*** toumi01 has quit IRC | 05:06 | |
*** Mud has joined #ingres | 05:07 | |
Dejan | wb Mud | 05:10 |
*** cthibert has joined #ingres | 05:11 | |
*** zxiiro has quit IRC | 05:13 | |
*** toumi01 has joined #ingres | 05:15 | |
*** KermitTheFragger has joined #ingres | 05:19 | |
KermitTheFragger | hi all | 05:19 |
KermitTheFragger | does anyone have experience with storing a large number of BLOB's in ingres ? | 05:19 |
KermitTheFragger | say 50.000 blob's of 200KB | 05:19 |
KermitTheFragger | would that be a good idea to do with ingres ? | 05:20 |
pboro | pretty often it is said that it's not a good idea | 05:24 |
pboro | because it is so slow to fetch them... if you store them in the db, implement some kind of caching in the application | 05:25 |
pboro | at least if they are accessed often | 05:25 |
KermitTheFragger | yeah i was afraid of that :) | 05:25 |
pboro | pretty usual way is to store them on the fs level and just store in the db a path and filename or similar... | 05:25 |
KermitTheFragger | yeah i know | 05:26 |
grantc | i don't think blob storage/retrieval is that slow | 05:26 |
KermitTheFragger | but then i have to ensure the integrity | 05:26 |
KermitTheFragger | also i have to make some mechanism to be able to hotbackup it all from the FS | 05:26 |
KermitTheFragger | replication | 05:26 |
KermitTheFragger | etc. | 05:26 |
pboro | grantc, I have tried it and without a caching layer it was awfully slow :( | 05:26 |
grantc | hmm | 05:26 |
pboro | I decided to leave them in the db anyway 'cos I don't want to break the data in pieces :/ | 05:27 |
KermitTheFragger | pboro: exactly my issue :) | 05:27 |
pboro | and the caching solution was pretty simple to implement on the application level | 05:27 |
grantc | fair enough - i did some tests whilst at CA for their content management system the performance was ok - not brilliant but ok | 05:27 |
pboro | grantc, how many requests per sec? | 05:28 |
pboro | and what was the size? :/ it depends on so many factors | 05:28 |
grantc | cannot remember, it was 6+ years ago | 05:28 |
pboro | yup | 05:28 |
KermitTheFragger | ok performance is also good enough for me on this one. The trouble of implementing the integrity constraints, hotbackup facilities, replication etc. is just not worth it | 05:29 |
grantc | varying - since it was a content management system i was playing with small and large files - mainly my mp3 collection iirc | 05:29 |
pboro | hehe | 05:29 |
KermitTheFragger | lol | 05:29 |
grantc | 200KB files should not be an issue | 05:29 |
Dejan | blobs are stored in pages, right ? | 05:29 |
grantc | bigger files might be | 05:29 |
Dejan | brrrr | 05:29 |
pboro | Dejan, they are... split into 2k or 8k or what ever pieces | 05:30 |
grantc | lobs are stored in separate tables | 05:30 |
pboro | lots of overhead :) | 05:30 |
Dejan | imagine mp3 file across 8k pages... | 05:30 |
Dejan | eeeek! | 05:30 |
grantc | i think you can define the default page sizes for those tables | 05:30 |
Dejan | or even smaller pages... | 05:30 |
Dejan | ingres should somehow just give a stream | 05:30 |
pboro | grantc, yup | 05:30 |
grantc | it's a trade off - speed vs integrity | 05:31 |
Dejan | and pipe some file somewhere in ii_blobs | 05:31 |
Dejan | (i know there is no ii_blobs) | 05:31 |
Dejan | (just an idea) | 05:31 |
Dejan | so, for website, one would change ii_blobs location | 05:32 |
Dejan | to somewhere visible on the web | 05:32 |
Dejan | (read only) | 05:32 |
Dejan | and basically can give a direct link | 05:32 |
pboro | blobs reminded me that I should build a test case for cache_dynamic and blobs, there's still some concurrency problem with them | 05:33 |
Dejan | guys, is there any other way to get start/end JNL and DMP file(s) except from infodb ? | 05:58 |
Dejan | parsing infodb output is a pain in the a** | 05:59 |
*** PaulM05_lunch is now known as PaulM05 | 06:07 | |
*** cthibert has quit IRC | 06:09 | |
Dejan | we have a problem on forum | 06:13 |
Dejan | many spammers | 06:14 |
Dejan | example: http://community.ingres.com/forum/dba-forum/51-error-crash.html | 06:14 |
*** zxiiro has joined #ingres | 06:14 | |
*** ChanServ sets mode: +v zxiiro | 06:14 | |
Dejan | similar: http://community.ingres.com/forum/migration-forum/11844-ingres-migration-tool-set.html | 06:15 |
KermitTheFragger | w00t! free nude pictures! Way better then all that boring talk about RDBM's ;-) | 06:17 |
*** mull has joined #ingres | 06:23 | |
*** Alex| has quit IRC | 06:50 | |
*** Dejan_ has joined #ingres | 06:52 | |
*** Dejan has quit IRC | 06:53 | |
*** Dejan_ is now known as Dejan | 06:57 | |
*** Alex| has joined #ingres | 07:04 | |
*** ChanServ sets mode: +o Alex| | 07:04 | |
*** DejanTest has joined #ingres | 07:35 | |
DejanTest | this is cool! | 07:36 |
DejanTest | http://webchat.freenode.net/ | 07:36 |
*** DejanTest has quit IRC | 07:37 | |
*** paulm_test has joined #ingres | 07:38 | |
paulm_test | :) | 07:38 |
Dejan | i did not know about this | 07:38 |
Dejan | i use ssh tunnel for irc ;) | 07:38 |
PaulM05 | I used to use mibbit.com before grant set up a proxy | 07:39 |
*** paulm_test has quit IRC | 07:39 | |
Dejan | i wonder if vectorwise uses opencl or similar ? | 07:44 |
Dejan | behind scenes | 07:44 |
Dejan | would it help improve performance ? | 07:44 |
Dejan | i bet some things in the optimiser can be parallelised | 07:45 |
*** cthibert has joined #ingres | 07:49 | |
*** Mud has quit IRC | 09:02 | |
*** Mud has joined #ingres | 09:08 | |
*** PaulM05 has quit IRC | 09:32 | |
*** Mud has quit IRC | 09:34 | |
*** Alex____ has joined #ingres | 09:38 | |
*** Alex| has quit IRC | 09:38 | |
*** Alex____ has quit IRC | 09:39 | |
*** Alex| has joined #ingres | 09:39 | |
*** ChanServ sets mode: +o Alex| | 09:39 | |
grantc | Dejan, regarding the spam - put up a post about it in the forums. We've been having an internal discussion about it but it would be useful to get an external voice in the mix | 09:49 |
Dejan | yes, but where ? | 09:50 |
grantc | http://community.ingres.com/forum/ingres-community/ would be a good place to start | 09:51 |
*** Alex| has quit IRC | 09:56 | |
*** KermitTheFragger has quit IRC | 10:06 | |
*** clach04 has joined #ingres | 10:45 | |
Dejan | hello clach04 | 10:45 |
clach04 | evenin' | 10:48 |
clach04 | well, I'm assuming it is "evening", I think you're in the UK Dejan :-) | 10:49 |
Dejan | it is | 10:49 |
Dejan | you are correct :D | 10:50 |
clach04 | morning over in on the West US coast | 10:50 |
Dejan | time to go home | 11:00 |
Dejan | in ~1 a football game is starting... | 11:00 |
Dejan | important one (for me) :) | 11:00 |
Dejan | Wolfsburg vs Fulham | 11:00 |
Dejan | bye all | 11:01 |
*** Dejan has quit IRC | 11:01 | |
*** DerMeister has joined #ingres | 12:34 | |
*** cthibert has left #ingres | 13:00 | |
*** DerMeister has quit IRC | 13:10 | |
*** zxiiro has quit IRC | 13:40 | |
*** grantc has quit IRC | 13:52 | |
*** bonro011 has joined #ingres | 13:58 | |
*** bonro01 has quit IRC | 14:00 | |
*** rossand has quit IRC | 14:40 | |
*** mull has quit IRC | 15:31 | |
*** atrofast has quit IRC | 15:47 | |
*** atrofast has joined #ingres | 15:55 | |
*** zxiiro has joined #ingres | 16:04 | |
*** ChanServ sets mode: +v zxiiro | 16:04 | |
*** atrofast has quit IRC | 18:26 | |
*** zxiiro has quit IRC | 18:30 | |
*** zxiiro has joined #ingres | 18:31 | |
*** ChanServ sets mode: +v zxiiro | 18:31 | |
*** epoitras has joined #ingres | 18:35 | |
*** cytrinox_ has joined #ingres | 19:21 | |
*** cytrinox has quit IRC | 19:24 | |
*** cytrinox_ is now known as cytrinox | 19:24 | |
*** epoitras has quit IRC | 20:39 | |
*** epoitras has joined #ingres | 20:48 | |
*** epoitras has quit IRC | 22:12 | |
*** Alex| has joined #ingres | 22:36 | |
*** ChanServ sets mode: +o Alex| | 22:36 | |
*** Alex| has quit IRC | 22:57 | |
*** Alex| has joined #ingres | 23:15 | |
*** ChanServ sets mode: +o Alex| | 23:15 | |
*** stliu has joined #ingres | 23:35 | |
stliu | withdefault, Ray? | 23:38 |
withdefault | Hi stliu | 23:38 |
stliu | hi, I think i commented on the wrong issue :( | 23:39 |
stliu | it should be HHH-4938 | 23:39 |
stliu | withdefault, would you please run FooBarTest again to see if the error in HHH-4938 also exist? | 23:39 |
stliu | it will faster than my connection :) | 23:40 |
withdefault | do I need to refresh any code? | 23:40 |
withdefault | in trunk or in 3.5.0-Final? | 23:41 |
stliu | in trunk, and do svn up | 23:41 |
stliu | i just commited the fix | 23:41 |
*** withdefault has quit IRC | 23:47 | |
*** Mud has joined #ingres | 23:48 | |
*** grantc has joined #ingres | 23:51 | |
*** withdefault has joined #ingres | 23:52 | |
*** ChanServ sets mode: +o withdefault | 23:52 | |
withdefault | stliu, sorry - network issues | 23:52 |
withdefault | stliu, I've been trying to debug through the code - where did you make the change? | 23:53 |
stliu | withdefault, <property name="str" column="`str`"/> in Baz.hbm.xml in test/legacy | 23:54 |
stliu | str is a function name in IngresDialect | 23:54 |
pboro | someone fixing the hibernate dialect? :) | 23:55 |
withdefault | ahh, I was looking in a completely different direction | 23:55 |
withdefault | hi pboro - yes, the dialect is being fixed and extended, but mostly fixed at the moment | 23:56 |
pboro | great! :) | 23:56 |
withdefault | stliu, you won't be able to get access to service desk yet - I need to get some things set up | 23:57 |
stliu | okay, thanks | 23:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!