| *** withdefault has joined #ingres | 00:57 | |
| *** ChanServ sets mode: +o withdefault | 00:57 | |
| *** withdefault sets mode: -o withdefault | 00:57 | |
| *** Mud has quit IRC | 00:58 | |
| *** Mud has joined #ingres | 01:04 | |
| *** KermitTheFragger has joined #ingres | 01:05 | |
| *** Dejan has joined #ingres | 01:59 | |
| Dejan | hello everybody | 01:59 |
|---|---|---|
| withdefault | hello | 02:00 |
| Dejan | guys, does Ingres terminal monitor have something like \G in MySQL/MariaDB ? | 02:23 |
| Dejan | where rows are displayed vertically | 02:23 |
| Dejan | i really need that functionality... | 02:23 |
| Dejan | are new admin tools already in ingres 10 ? | 03:25 |
| *** Dejan has quit IRC | 03:51 | |
| *** Dejan_ has joined #ingres | 03:57 | |
| *** Dejan_ has joined #ingres | 03:59 | |
| *** Dejan_ has joined #ingres | 03:59 | |
| *** Dejan_ is now known as Dejan | 04:01 | |
| Dejan | Finally i wasn't lazy to make a proper tunelling - so I can use X-Chat for IRC, from my office :) | 04:03 |
| *** cthibert has joined #ingres | 04:04 | |
| *** Mud has quit IRC | 04:59 | |
| * Dejan is going to have a lunch | 05:00 | |
| *** Mud has joined #ingres | 05:04 | |
| *** Mud|lalal has joined #ingres | 05:13 | |
| *** Mud has quit IRC | 05:13 | |
| *** zxiiro has quit IRC | 05:25 | |
| * Dejan is back | 05:33 | |
| withdefault | Dejan, why are you looking for terminal monitor output like MySQL \G? | 05:41 |
| Dejan | because when you have a table with 20 columns you cannot see what is what | 05:41 |
| Dejan | :D | 05:41 |
| Dejan | simple as that | 05:42 |
| withdefault | ok, fair point - I usually use something other than tm | 05:43 |
| Dejan | i mostly use TM | 05:47 |
| Dejan | (sql actually) | 05:47 |
| Dejan | but SQL just calls TM :D | 05:47 |
| Dejan | ans Squirrel-SQL | 05:47 |
| Dejan | and* | 05:47 |
| *** DarylM has joined #ingres | 05:53 | |
| Dejan | hiya DarylM | 06:06 |
| DarylM | Good day. Glad to be back (from college visits....) | 06:06 |
| *** zxiiro has joined #ingres | 06:11 | |
| *** ChanServ sets mode: +v zxiiro | 06:11 | |
| *** rossand has joined #ingres | 06:37 | |
| *** ChanServ sets mode: +o rossand | 06:37 | |
| *** cthibert has left #ingres | 06:41 | |
| Dejan | question - if i modify a table to reconstruct, and different fillfactor - does it immediately shows? | 07:01 |
| Dejan | for some reason i still see the same number of overflow pages | 07:02 |
| Dejan | after i modified the table | 07:02 |
| Dejan | or i need to do something else | 07:02 |
| Dejan | ? | 07:02 |
| *** cthibert has joined #ingres | 07:07 | |
| *** PaulM05 has joined #ingres | 07:10 | |
| Dejan | hey paul :) | 07:11 |
| PaulM05 | hi | 07:11 |
| Dejan | how's going ? | 07:11 |
| PaulM05 | not bad - hopefully have a fix for my latest issue - final stages of testing now | 07:12 |
| PaulM05 | Dejan - have you used isql? | 07:15 |
| Dejan | i prefer sql | 07:20 |
| Dejan | sql with -history_recall :D | 07:21 |
| PaulM05 | fair enough | 07:21 |
| Dejan | Paul, short question - when i modify table | 07:21 |
| Dejan | will Ingres change it immediately ? | 07:21 |
| PaulM05 | just looking at the IRC logs - that \G thing doesn't exist right now - but it's a nice small(ish) open source project | 07:21 |
| PaulM05 | yes | 07:22 |
| Dejan | because what i did - i modified a HASH table | 07:22 |
| Dejan | that had 5000 overflow pages | 07:22 |
| Dejan | to reconstruct, with fillfactor 60% | 07:22 |
| Dejan | and when i queried iitables | 07:22 |
| Dejan | it still shows those 5000 overflow pages | 07:22 |
| Dejan | for that particular table | 07:22 |
| PaulM05 | did you specify minpages? | 07:25 |
| Dejan | i think i did not | 07:25 |
| Dejan | i did not create that table... | 07:25 |
| Dejan | it is some legacy crap... | 07:26 |
| PaulM05 | you can change minpages when you modify | 07:26 |
| PaulM05 | I usually increase minpages when trying to get rid of hash overflow | 07:27 |
| Dejan | it is set to 10... | 07:28 |
| Dejan | :D | 07:28 |
| Dejan | and it has 25k allocated pages... | 07:28 |
| PaulM05 | hmm - probably going to get a *few* key collisions then ;) | 07:28 |
| Dejan | :)))))))))))))))) | 07:28 |
| Dejan | i thought when i specify fillfactor, ingres automatically finds out how many min/max pages to set | 07:31 |
| Dejan | judging how many rows i have in the table, row widths etc | 07:32 |
| PaulM05 | I don't know for sure but based on my experience I'd say it probably doesn't | 07:32 |
| Dejan | strange, now i changed minpages | 07:40 |
| PaulM05 | still showing overflow? | 07:40 |
| Dejan | whatever i set for minpages, it always shows the same number of overflow pages | 07:41 |
| Dejan | :D | 07:41 |
| Dejan | looks like it does not change anything | 07:41 |
| Dejan | by default it changes maxpages to like 80k pages | 07:42 |
| PaulM05 | are you commit-ing after the modify? Shouldn't need to to see the change but it might make a difference | 07:42 |
| Dejan | no, 800k pages | 07:42 |
| Dejan | i did commit | 07:42 |
| Dejan | weird... | 07:44 |
| Dejan | really, whatever i do - it does not show change in number of overflow pages... | 07:45 |
| PaulM05 | you could look at the underlying files to see if they've changed size | 07:45 |
| PaulM05 | do you know about iifile_info etc? | 07:45 |
| Dejan | sure | 07:45 |
| Dejan | i thought iitables is immediately updated... | 07:45 |
| Dejan | IMHO this is a bug | 07:45 |
| PaulM05 | have you looked at what the distribution of keys is? | 07:46 |
| Dejan | for this particular table i do not have keys | 07:46 |
| Dejan | it is just a simple backup stuff | 07:46 |
| PaulM05 | I mean the column(s) it is hash on | 07:46 |
| PaulM05 | physical key | 07:47 |
| Dejan | no, underlying file did not change | 07:51 |
| Dejan | at all | 07:51 |
| PaulM05 | that is strange | 07:51 |
| Dejan | ok, i will now SHA1 sum it | 07:52 |
| *** mull has joined #ingres | 07:56 | |
| Dejan | yeah, it is not changed | 07:57 |
| Dejan | could it be in cache ? | 07:57 |
| Dejan | pretty much impossible... | 07:58 |
| PaulM05 | how big is it? | 07:58 |
| Dejan | the file is 9M | 07:58 |
| Dejan | still, iitables should be changed... | 07:58 |
| Dejan | even if it is in cache | 07:58 |
| Dejan | i am really buffled | 07:58 |
| PaulM05 | me too | 07:59 |
| Dejan | funny thing is - iitables shows modification in minpages, maxpages... | 08:04 |
| Dejan | but does not change overflow pages | 08:04 |
| Dejan | no matter what i dod | 08:05 |
| Dejan | do* | 08:05 |
| Dejan | which is impossible | 08:05 |
| Dejan | that number should change, if i understand how ingres works :) | 08:05 |
| PaulM05 | thing is those values are stored but overflow pages is read directly from the file header | 08:05 |
| PaulM05 | so it's even more impossible ;) | 08:05 |
| Dejan | i checked with sha1sum - file is not changed | 08:06 |
| withdefault | hi rossand, atrofast - is there an automated way to create svn repository folders for new community projects? or is it a manual process? | 08:06 |
| atrofast | withdefault: It's pretty much a manual process | 08:07 |
| withdefault | who should I direct requests to :) | 08:07 |
| rossand | withdefault: just about to say ;-) We need to set up some access rules for the project, LDAP. | 08:08 |
| atrofast | If you email me the name of the new branch you'd like and your user-name I'll set it up for you | 08:08 |
| PaulM05 | Dejan - try running trace point 1314 - will force an archiver refresh and therefore a consistency point | 08:08 |
| withdefault | ok, thanks - expect something via your email | 08:08 |
| atrofast | No problem :) | 08:08 |
| PaulM05 | DM1314 that should be | 08:09 |
| Dejan | is it safe to run on a production server ? | 08:11 |
| PaulM05 | yes | 08:12 |
| withdefault | thanks rossand, atrofast - have passed the information along | 08:12 |
| Dejan | PaulM05, no change in file | 08:15 |
| Dejan | after the tracepoint | 08:15 |
| Dejan | i will investigate this later at home... | 08:16 |
| Dejan | i am getting tired... | 08:16 |
| Dejan | it must be something i am overlooking :D | 08:16 |
| PaulM05 | my next suggestion would be to try modifying to heap then back to hash | 08:16 |
| Dejan | ok, will do that :) | 08:16 |
| PaulM05 | if it doesn't change after that - well by then I'll be home ... :) | 08:17 |
| Dejan | it changed | 08:18 |
| Dejan | khm... when i modified it to HASH, again completely the same as prefiously :D, 270 overflow pages... | 08:22 |
| Dejan | this is interesting | 08:22 |
| Dejan | maybe settings i gave are not good :D | 08:22 |
| PaulM05 | maybe you just have lots of the same key values | 08:23 |
| Dejan | ok, i learned something | 08:27 |
| Dejan | PaulM05, i have found where was my mistake | 08:27 |
| Dejan | see, as this table did not have any keys set | 08:27 |
| PaulM05 | yes? | 08:28 |
| *** Mud|lalal has quit IRC | 08:28 | |
| Dejan | i thought i should not hash it on anything | 08:28 |
| Dejan | so i just used MODIFY tablename TO HASH WITH ... | 08:28 |
| Dejan | did not specify what | 08:28 |
| Dejan | thus i always got the same number of overflow pages i guess | 08:28 |
| PaulM05 | what did it do - first column? | 08:28 |
| Dejan | i thought ingres takes care of what to hash in that case :D | 08:29 |
| Dejan | but what i did now - i analysed the table | 08:29 |
| Dejan | and found what columns could be used as good candidates for a key | 08:29 |
| Dejan | and hashed on 3 columns | 08:29 |
| Dejan | (because table contains duplicate data...) | 08:29 |
| Dejan | and from 8900 overflow pages i came to 29 | 08:30 |
| Dejan | nice, now table takes 2.2M | 08:31 |
| Dejan | from 9M to 2.2M | 08:31 |
| Dejan | brilliant | 08:31 |
| PaulM05 | what was it using for the hash before? the first column? | 08:31 |
| Dejan | nothing | 08:31 |
| Dejan | :) | 08:31 |
| Dejan | so i suppose yes, first column | 08:32 |
| Dejan | and that column has maaaaaany duplicates | 08:32 |
| PaulM05 | it used to be that if you didn't specify a column it would give an error - I suspect now it will use a primary/unique key if there is one otherwise the first column | 08:33 |
| Dejan | so... CREATE TABLE tablename (tbl_id ... PRIMARY KEY, ...) WITH STORAGE=HASH will efectively HASH on that PK specified? | 08:36 |
| PaulM05 | possibly - I'm partly guessing - let me look it up | 08:37 |
| Dejan | what i did so far is - created table, MODIFY-ed it to HASH, keyed on a PK, then made a PK that uses storage structure's key | 08:39 |
| *** bonro011 has quit IRC | 08:39 | |
| *** bonro01 has joined #ingres | 08:39 | |
| Dejan | hehe, a question that i never asked myself before - how to find out what is the column underlying HASH structure uses? | 08:40 |
| Dejan | let me find out | 08:41 |
| PaulM05 | help table | 08:41 |
| Dejan | hell i hate when it makes a BTREE index for a primary key | 08:46 |
| Dejan | even though i have a hash table | 08:46 |
| Dejan | keyed on that column | 08:46 |
| Dejan | :) | 08:46 |
| PaulM05 | did you use "WITH INDEX=BASE TABLE STRUCTURE"? | 08:46 |
| Dejan | no, because i do not know how to specify what column is going to be the key for HASH structure | 08:47 |
| PaulM05 | MODIFY mytable TO HASH ON col1(,col2,...) | 08:48 |
| Dejan | sure, i know that | 08:48 |
| Dejan | but how to do that in the create statement ? | 08:49 |
| Dejan | erm, is it possible at all ? | 08:49 |
| PaulM05 | are you doing a CREATE TABLE AS SELECT? | 08:50 |
| Dejan | nope, i am just playing, and want to do as much i can in one single statement :) | 08:51 |
| Dejan | i guess i want too much :) | 08:52 |
| PaulM05 | have a look at table_autostructure in CBF | 08:53 |
| PaulM05 | something else to play with (and find bugs) | 08:53 |
| Dejan | guys, is it possible to have connection information in the auditdb ? | 08:59 |
| Dejan | i want do see who did particular transaction | 09:00 |
| Dejan | and possibly from which host | 09:00 |
| PaulM05 | that's not in the files - just the user name | 09:00 |
| Dejan | how to get username ? | 09:01 |
| PaulM05 | should be in the BEGIN record isn't it? | 09:02 |
| Dejan | we have a legacy (OpenVMS) application which connects to our new Ingres server | 09:05 |
| Dejan | it always connects as a specific user | 09:05 |
| Dejan | however, when i examine current sessions | 09:05 |
| Dejan | i can see their username | 09:05 |
| Dejan | (because on vms they connect using their own usernames) | 09:05 |
| Dejan | but in auditdb i always get one, single user :( | 09:06 |
| Dejan | recently someone deleted some information using that application | 09:06 |
| Dejan | and we wanted to find out who | 09:06 |
| PaulM05 | only the user on the DBMS side is stored in the journal file. The client side info isn't part of that. | 09:07 |
| Dejan | but information who did cannot be extracted from auditdb | 09:07 |
| Dejan | yep | 09:07 |
| Dejan | :( | 09:07 |
| Dejan | is there any way to enable something | 09:07 |
| Dejan | on VMS side | 09:07 |
| Dejan | so i can later on pinpoint user ? | 09:07 |
| PaulM05 | for the future? yes | 09:08 |
| Dejan | really? how? | 09:08 |
| PaulM05 | actually - I was thinking of security auditing but that doesn't log the client information either | 09:10 |
| PaulM05 | can you change the legacy app? | 09:11 |
| Dejan | yes | 09:11 |
| Dejan | we can store info | 09:11 |
| Dejan | about session | 09:11 |
| Dejan | in some table | 09:11 |
| Dejan | that was the first thing that came to our mind | 09:11 |
| PaulM05 | well you could change the connection to use installation password and then the client user would be the DBMS user | 09:12 |
| PaulM05 | of course you might need to set those users up | 09:12 |
| Dejan | you mean that VMS app uses "ingres" user to connect ? | 09:13 |
| Dejan | pardon, should use | 09:13 |
| PaulM05 | no | 09:13 |
| PaulM05 | at the moment it uses a vnode with a single user name in it correct? so it always appears as the same user | 09:13 |
| Dejan | yes | 09:13 |
| PaulM05 | so you can set up the vnode so it uses a special password called the installation password and instead of a username you put "*" | 09:14 |
| PaulM05 | then when it makes the connection it uses the local client user as the target user | 09:14 |
| PaulM05 | which must exist in the target installation | 09:15 |
| Dejan | khm... | 09:15 |
| Dejan | i have all users migrated to the new server | 09:15 |
| Dejan | so basically both servers have them all | 09:15 |
| Dejan | i will try that, thanks | 09:15 |
| Dejan | a lot | 09:15 |
| Dejan | :D | 09:15 |
| PaulM05 | have a look at the connectivity guide and look for "installation passwords" | 09:15 |
| PaulM05 | I need to go - have a good Easter everyone! | 09:19 |
| Dejan | You too! | 09:19 |
| *** PaulM05 has quit IRC | 09:19 | |
| *** Dejan has quit IRC | 09:46 | |
| *** KermitTheFragger has quit IRC | 10:00 | |
| *** withdefault has quit IRC | 10:32 | |
| *** zxiiro has quit IRC | 10:55 | |
| *** DerMeister has joined #ingres | 11:00 | |
| *** cthibert has left #ingres | 11:40 | |
| *** epoitras has joined #ingres | 11:45 | |
| *** epoitras has quit IRC | 11:46 | |
| *** DerMeister has quit IRC | 12:27 | |
| *** rossand has quit IRC | 13:13 | |
| *** rossand has joined #ingres | 13:26 | |
| *** ChanServ sets mode: +o rossand | 13:26 | |
| *** rossand has quit IRC | 13:27 | |
| *** rossand has joined #ingres | 14:36 | |
| *** ChanServ sets mode: +o rossand | 14:36 | |
| *** rossand has quit IRC | 14:36 | |
| *** zxiiro has joined #ingres | 15:13 | |
| *** ChanServ sets mode: +v zxiiro | 15:13 | |
| *** mull has quit IRC | 16:03 | |
| *** epoitras has joined #ingres | 16:28 | |
| *** rossand has joined #ingres | 19:18 | |
| *** ChanServ sets mode: +o rossand | 19:18 | |
| *** cytrinox has quit IRC | 19:24 | |
| *** cytrinox has joined #ingres | 19:26 | |
| *** Eddie has joined #ingres | 19:49 | |
| *** Eddie is now known as Guest65113 | 19:49 | |
| *** epoitras has quit IRC | 19:52 | |
| *** Guest65113 is now known as epoitras | 20:02 | |
| *** rossand has quit IRC | 23:22 | |
| *** Mud|lalal has joined #ingres | 23:34 | |
| *** zxiiro has quit IRC | 23:50 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!