*** dyki has joined #ingres | 00:49 | |
Vroomfondle | is there some kind of general changelog which stretches back to Ingres 2.6? My boss is curious as to exactly what advantages we'll see when we move to 9.2 from 2.6 | 01:41 |
---|---|---|
grantc | Vroomfondle: there is the migration guide which has the info i believe | 01:52 |
grantc | http://docs.ingres.com/Ingres/9.2/Migration%20Guide/docindex.htm. | 01:52 |
Vroomfondle | thanks, grantc | 01:59 |
*** grantc_ has joined #ingres | 02:00 | |
*** ChanServ sets mode: +o grantc_ | 02:00 | |
*** grantc_ has quit IRC | 02:00 | |
*** withdefault has joined #ingres | 03:35 | |
*** ChanServ sets mode: +o withdefault | 03:35 | |
*** [1]Gerhard has joined #ingres | 04:26 | |
[1]Gerhard | Hi all | 04:26 |
*** [1]Gerhard has quit IRC | 04:27 | |
*** Dejan has joined #ingres | 04:37 | |
Dejan | hi guys | 04:38 |
*** cthibert has joined #ingres | 05:42 | |
Dejan | i wander if there is any way to speedup inserts ... | 06:24 |
Dejan | wonder* | 06:24 |
Vroomfondle | remove any unneccessary indexes | 06:30 |
Dejan | i have no secondary indices | 06:40 |
Dejan | well at least i think i do not | 06:40 |
Dejan | on average i have around 500 inserts pers second | 06:41 |
Dejan | i cannot make a file and use COPY because I am processing each row that needs to be inserted | 06:42 |
pboro | Dejan, is your transaction log located on a different disk than data? are the tables journalled? | 06:44 |
Dejan | pboro: TL is not located on a different disk | 06:50 |
Dejan | which is a good idea | 06:50 |
Dejan | did not think about it | 06:50 |
Dejan | because i am still doing the migration stuff here | 06:50 |
Dejan | so more/less i am testing our software | 06:50 |
Dejan | i am definitely going to put logs on a different disk | 06:51 |
Dejan | it is all on virtualised server(s) | 06:52 |
pboro | I have a DB server with 3x146GB mirrors, I have Ingres spread between different mirrors based on the ingres documentation | 06:52 |
Dejan | i have enough ram on this server to have whole DB in it | 06:53 |
Dejan | however, ingres does not have that feature... | 06:53 |
pboro | your OS will cache the files in memory if you have enough memory anyway | 06:54 |
pboro | so I don't see a need for such a feature, other option is to make DMF caches huge so that the data fits them | 06:54 |
Dejan | not sure... | 06:54 |
Dejan | i have ~500 tables | 06:55 |
Dejan | OS will cache those frequently accessed | 06:55 |
Dejan | which will help | 06:55 |
pboro | the number of tables is not critical, the amount of data is | 06:55 |
Dejan | i must read docs again :) | 07:00 |
Dejan | do not remember reading about good practices when it comes to settings up ingres server | 07:00 |
pboro | Dejan, here's one: http://docs.ingres.com/Ingres/9.2/Installation%20Guide/samplediskconfigurations.htm | 07:01 |
Dejan | thank you, mate | 07:01 |
Dejan | aha | 07:02 |
Dejan | so you have logs where the CKPs are? | 07:02 |
Dejan | ah no | 07:03 |
Dejan | where II_SYSTEM is | 07:03 |
Dejan | ok | 07:03 |
Dejan | makes sense | 07:03 |
pboro | also I only use 15k rpm disks | 07:04 |
Dejan | yep | 07:04 |
Dejan | that is what we have here | 07:04 |
Dejan | see, it is only during the migration process when we are going to have zillion inserts | 07:05 |
Dejan | after that it is going to be a "normal" database | 07:05 |
Dejan | :D | 07:05 |
Dejan | guys, if want to change location of, ie. II_DATABASE | 07:45 |
Dejan | how do i do that? | 07:45 |
Dejan | can i just set the environment? | 07:45 |
Dejan | and restart ingres? | 07:45 |
grantc | Dejan: relocatedb | 07:57 |
grantc | the database locations are set in different locations | 07:58 |
*** FrankW has joined #ingres | 08:10 | |
Dejan | grantc: what if i want all databases to be moved to a new location? | 08:25 |
Dejan | and also future ones | 08:25 |
grantc | you will have to move each db individually for existing dbs. In the second case there are flags for createdb to define the locations | 08:27 |
Dejan | so there is no "from now on use /path/to/dir location for new databases" ? | 08:35 |
Dejan | it is hard to believe | 08:36 |
grantc | not that i am aware of | 08:50 |
monda07_ | Ingres has hardcoded path names in the aaaaaaaa.cnf file. It is a very old legacy and making that change would be more effort than the benefit (or at least that is the default explanation) | 08:52 |
*** atrofast1 has joined #ingres | 09:47 | |
*** atrofast has quit IRC | 09:50 | |
*** grantc has quit IRC | 09:51 | |
*** cthibert has left #ingres | 09:58 | |
*** toumi01 has quit IRC | 10:07 | |
*** Dejan has quit IRC | 11:08 | |
*** cthibert has joined #ingres | 11:35 | |
*** rossand1 has joined #ingres | 12:05 | |
*** rossand1 has quit IRC | 12:18 | |
*** FrankW has quit IRC | 12:18 | |
*** rossand1 has joined #ingres | 12:18 | |
*** FrankW has joined #ingres | 12:18 | |
*** rossand has quit IRC | 12:19 | |
*** DerMeister has joined #ingres | 12:22 | |
*** bonro011 has left #ingres | 13:55 | |
*** bonro01 has joined #ingres | 13:56 | |
*** cthibert has left #ingres | 14:03 | |
*** withdefault has quit IRC | 14:07 | |
*** DerMeister has quit IRC | 14:26 | |
*** rossand1 has quit IRC | 14:26 | |
*** FrankW has quit IRC | 14:26 | |
*** NoeJeko has quit IRC | 14:26 | |
*** NoeJeko has joined #ingres | 14:26 | |
*** DerMeister has joined #ingres | 14:26 | |
*** rossand1 has joined #ingres | 14:26 | |
*** FrankW has joined #ingres | 14:26 | |
*** monda07_ has quit IRC | 14:55 | |
*** DerMeister has quit IRC | 15:04 | |
*** rossand1 has quit IRC | 16:59 | |
*** clach04 has quit IRC | 17:36 | |
*** ii_log has joined #ingres | 20:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!