Time |
S |
Nick |
Message |
11:04 |
|
kados |
morning all |
11:21 |
|
owen |
Hi kados |
11:21 |
|
owen |
's front walk |
18:47 |
|
teknotus |
Hi anybody here? |
19:04 |
|
thd |
teknotus: what help do you need? |
20:32 |
|
teknotus |
thd: I installed the latest koha on a debian sarge system last week, and it mostly worked, but I eventually noticed some things that weren't, and tracked them down to sql failures, but didn't figure out what was being sent to the server. |
20:33 |
|
teknotus |
So my guess is that the version of MySQL that comes with sarge isn't compatible with the latest koha |
20:33 |
|
teknotus |
And I wanted to know if what was the best solution |
20:33 |
|
thd |
teknotus: What version of MySQL are you using? |
20:38 |
|
teknotus |
Hmm I'm not sure. That machine is 20 miles away, and behind a firewall. I would assume that since I have a debian sarge system too, and it tells me that the version it wants to install is 4.0.24 that the system with koha on it has that version too. |
20:39 |
|
thd |
teknotus: What is the result of mysql -V? |
20:40 |
|
teknotus |
I don't have it installed on this computer it's on another computer far away. |
20:40 |
|
teknotus |
I guess I could install it right now. |
20:41 |
|
thd |
teknotus: Do you have ssh access to the machine 20 miles away? |
20:41 |
|
teknotus |
Nope, that's why I mentioned it was behind a firewall, and it is highly unlikely that anyone will punch a hole in it for me. |
20:42 |
|
thd |
teknotus: security can be painful :) |
20:42 |
|
teknotus |
Does 4.0.24 sound like an unlikely version? |
20:43 |
|
thd |
teknotus: There is a small problem with some version of MySQL 4 on Koha |
20:45 |
|
thd |
teknotus: That may be your version but upgraded all my systems with MySQL to Debian testing (etch). |
20:46 |
|
teknotus |
I was just setting up a test system to try it out. The plan is to eventually turn it over to the administrators to install onto one of the web servers with hole punched in the firewall so that people can get to the OPAC. |
20:46 |
|
thd |
teknotus: The Koha installer fails to grant some type of privilege that is new with MySQL 4.X to the Koha user. |
20:47 |
|
teknotus |
And I'm concerned that if I am requiring a custom version of MySQL they will be less likely to actually get it installed. |
20:47 |
|
thd |
teknotus: That problem will not cause Koha to fail completely but will give error messages in the Koha error log.. |
20:49 |
|
teknotus |
Well the symptom that eventually lead to me looking in the logs was when I was adding a book, and trying to get it to download a marc record from the library of congress based on the ISBN. I went to the LC own web interface, and verified that it was there first. |
20:49 |
|
teknotus |
But it wasn't finding anything in it's search |
20:49 |
|
teknotus |
I don't know if that helps |
20:50 |
|
teknotus |
Can it be fixed by manually running some SQL grant queries? |
20:51 |
|
thd |
teknotus: Simply grant the privilege to the Koha user that the error log complaint requires to satisfy and grant the privilege to the Koha user. Simply look for missing privileges out of possible grants. |
20:52 |
|
teknotus |
As I remember it it didn't specify what the SQL error was just that it failed. |
20:52 |
|
thd |
teknotus: Z39.50 problems for searching libraries are entirely different. |
20:52 |
|
teknotus |
In the log it didn't have any record of even trying to do a Z30.50 search |
20:53 |
|
teknotus |
But it had an SQL failure in the log at the time when I tried to do the search |
20:53 |
|
thd |
teknotus: Z39.50 requires a special daemon to run and that daemon has its own configuration file. |
20:54 |
|
teknotus |
So I thought that it was running an SQL query in the same script that was going to trigger the Z39.50 call, and when the SQL failed it never got to the Z39.50 part. |
20:54 |
|
teknotus |
Yeah I figured that part out. I had the daemon running. |
20:56 |
|
thd |
teknotus: The SQL problem may be unrelated but before experimenting with adding records you should verify that you have a valid System Administration : MARC Check. |
20:57 |
|
teknotus |
That failed on one count. |
20:58 |
|
thd |
teknotus: Then run rebuildnonmarc.pl after any MARC to Koha Link changes. |
20:58 |
|
thd |
teknotus: What was the failure? |
20:59 |
|
teknotus |
I don't remember what the failure was. Just that I couldn't find anything that explained what to do in that situation. |
20:59 |
|
thd |
teknotus: MARC check must have no errors to avoid data problems. |
21:00 |
|
teknotus |
If it fails with it refuse to do anything? |
21:00 |
|
teknotus |
Or silently corrupt things |
21:00 |
|
thd |
teknotus: If it fails you will have silent corruption of bibliographic data at least. |
21:01 |
|
teknotus |
Is this rebuildnonmark.pl a script I need to run from the command line, and can it fix the problem that is causing MARC Check to fail? |
21:02 |
|
thd |
teknotus: I have always purged and reloaded bibliographic data after an error on MARC Check. |
21:02 |
|
thd |
teknotus: Just to be safe. |
21:03 |
|
teknotus |
I never got any books into the system because I wanted to be able to have it automatically fetch MARC records when possible, and it was only going to be a test setup anyways. |
21:03 |
|
thd |
teknotus: The problem is that MARC 21 Koha comes with defaults that do not work immediately. |
21:06 |
|
teknotus |
Hmm, and I was initially thinking that it would be as simple as going back to an older version of koha that didn't use some new feature of the latest MySQL that breaks on an old version of MySQL. |
21:06 |
|
thd |
teknotus: If you have an email address, I can send you an email message that has information about things to check in order for the Z39.50 server to run well. |
21:07 |
|
thd |
teknotus: s/server/client/. |
21:07 |
|
teknotus |
Yeah just send to me at my nick at gmail.com |
21:08 |
|
thd |
teknotus: I will send you a message Monday as you have no access to the Koha machine until then I presume.. |
21:09 |
|
thd |
teknotus: I am writing a new Z39.50 client for Koha that will be much less troublesome. |
21:10 |
|
teknotus |
Send it at your conveniance. I don't know for sure that I can work on it until Tuesday, but with gmail it is always easy to search for an old message. |
21:11 |
|
thd |
teknotus: I am also working on some better defaults for MARC 21 Koha that should appear in 2.26 next month. |
21:12 |
|
teknotus |
I wrote a little Z39.50 client months ago. It just took a ISBN from the command line, and returned title author, and LC/Dewey info. |
21:14 |
|
teknotus |
I've been trying to help setup a small library for freegeek.org |
21:15 |
|
teknotus |
Libraries are far more complicated than I initially imagined |
21:15 |
|
thd |
teknotus: Z39.50 clients can be very simple. The one in Koha is probably the single greatest source of initial user problems with Koha. |
21:17 |
|
teknotus |
I have some appreciation of how complicated the protocal is. |
21:19 |
|
thd |
teknotus: The complexities of the MARC record format were grafted on to Koha. Koha 3 will use a textual database designed for library systems to keep MARC out of SQL. |
21:21 |
|
thd |
teknotus: Much of the complexity of Z39.50 cannot be widely used because often targets only support a small subset of what the standard allows. |
21:23 |
|
teknotus |
I forget. Does the standard allow you to query the server to find out what id does support? |
21:23 |
|
teknotus |
s/id/it/ |
21:23 |
|
teknotus |
Yeah I read about that the "index data" zebra product |
21:24 |
|
thd |
teknotus: Koha will also sometimes fail to obtain some records when querying multiple clients. That has been a problem with lack of documentation for querying multiple targets correctly in the underlying Perl module. |
21:24 |
|
teknotus |
another library system I was considering emilda.org uses that already I think. |
21:25 |
|
thd |
teknotus: Z39.50 explain is supposed to return what the server supports. |
21:27 |
|
teknotus |
I found that searching for open source library software was problematic because I was getting many hits for libraries of functions polluting my results |
21:27 |
|
teknotus |
And I didn't find koha for quite a while. |
21:27 |
|
teknotus |
I found emilda searching for MARC, and Z39.50 |
21:28 |
|
teknotus |
I only found out about koha via a tip |
21:34 |
|
thd |
teknotus: There is a site which lists about all open source library systems |
21:41 |
|
thd |
teknotus: It is not always well updated but http://www.oss4lib.org/ has been the best single source for identifying open source library projects for actual library management. |
21:45 |
|
thd |
teknotus: My own involvement with Koha aside from being a distant observer has been since mid last year as it seemed to have the most significant level of feature development from a larger number of developers. |
21:48 |
|
thd |
teknotus: Koha is the oldest OSS ILS so it has had a head start and still has the fastest development cycle and seems to have the most ambitious plans. |
21:50 |
|
teknotus |
Yeah |
21:51 |
|
teknotus |
The biggest design flaw seems to be the use of SQL to store marc records, but that is in the process of being fixed. |
21:53 |
|
teknotus |
It would also be really nice if everything was in debian packages, but I must say that the installer is slick, and in some ways more useful than having a .deb because it is easy to install multiple copies in whatever way you want. |
21:55 |
|
thd |
teknotus: The intentions of some for Koha are to compete and surpass every feature that proprietary ILS systems have as rapidly as may be practical. |
21:56 |
|
thd |
s/that/that even the most expensive/ |
22:01 |
|
thd |
teknotus: There has been a Debian package in the long past and will be again in some future but there is not enough time from those who have experience maintaining Debian packages. |
22:03 |
|
teknotus |
The documentation for creating debian packages is somewhat overwhelming. |
22:05 |
|
teknotus |
I think surpasing proprietary systems features is a typical free software project goal, that isn't always expressly stated. Linus said that it was never his intention to hurt Microsoft, just an unintended side effect. But for many of the people working on linux it is an explicit goal. |
22:05 |
|
thd |
teknotus: Someone with experience volunteered a colleague of his a few months ago but he seems not to have had time. |
22:07 |
|
teknotus |
Well I got as far as making 3 .deb support packages for perl libraries required by koha, but koha itself is more complicated. |
22:08 |
|
thd |
teknotus: It is not necessary to hurt proprietary software. It is better to Co opt proprietary software companies. Get them on your side.. |
22:09 |
|
teknotus |
Yeah I know |
22:10 |
|
teknotus |
Sometimes a proprietary software company releases a free software product, and people are stunned. |
22:12 |
|
thd |
teknotus: The Koha project leader has committed to putting the Koha libraries in their own CPAN namespace. Packaging should then be much easier. |
22:13 |
|
thd |
s/libraries/Perl modules/ |
22:14 |
|
teknotus |
Those are synonyms to me |
22:15 |
|
thd |
:) |
22:16 |
|
teknotus |
I program in both perl, and C |
22:16 |
|
teknotus |
Conceptually they are the same |
22:16 |
|
teknotus |
But different names for them. |
22:32 |
|
thd |
teknotus: Perl CPAN module support for working with MARC records tends to make Perl a natural language for library systems if for no other reason. |
23:32 |
|
thd |
kados: are you around? |
23:35 |
|
thd |
kados: I wanted to be using the most up to date rel_2_2 with whatever 2.6 features might be there as the starting point for creating a complete and accurate MARC 21 default bibliographic framework. |
23:37 |
|
thd |
kados: I have not started with the framework because of a strange problem with the OPAC in the latest rel_2_2 from cvs. |
23:39 |
|
thd |
kados: What are you using from CVS past 2.25 on your 2.25 test? |
05:24 |
|
thd |
kados: nevermind, problem solved |