Forum .LRN Q&A: Moving .LRN from server

Posted by Sergio Izquierdo on 09/09/10 03:41 PM

Hi, I am having a big problem moving from server.
I have a Fedora Server with Postgres 7.x and AOLServer 4. I have migrated the database, copied all the files of .LRN (changed the configuration), installed Posgres 7.y and AOLServer 4.

When I start the aolserver in the new Server (with SUSE), I can see the .LRN framework. I can add new courses, new information. But when I click to see an old course, I just hangs. And there is no more info displayed on the error.log file.

What could be wrong?

Thank you

Sergio

2: Re: Moving .LRN from server (response to 1)

Posted by Gustaf Neumann on 09/10/10 07:43 AM

is there maybe a redirect on that page to the "old machine"? Check the kernel parameter.

3: Re: Moving .LRN from server (response to 2)

Posted by Sergio Izquierdo on 09/10/10 02:42 PM

Well, not really, I also try to access with an admin user, going to the course section (where you can delete, edit, etc., the courses) and I can only access the new courses. When I access the old ones, it crashes. I am accesing all the server with an URL with the IP of the server.

I also had to change this when I was compiling the aolserver:
http://openacs.org/forums/message-view?message_id=309887
http://openacs.org/forums/message-view?message_id=1167801
http://openacs.org/forums/message-view?message_id=309847

Thanks for your help

Sergio

4: Re: Moving .LRN from server (response to 3)

Posted by Dave Bauer on 09/10/10 02:56 PM

Double check the stacksize, a crash with no logging usually indicates that the server ran out of stack space.

5: Re: Moving .LRN from server (response to 3)

Posted by Gustaf Neumann on 09/10/10 03:27 PM

if you are running into compilation errors with 64bit, then you are working with an old version of aolserver. get aolserver 4.5.1!

6: Re: Moving .LRN from server (response to 4)

Posted by Sergio Izquierdo on 09/10/10 05:10 PM

I made 5x the size of the stack and the same

7: Re: Moving .LRN from server (response to 5)

Posted by Sergio Izquierdo on 09/10/10 05:30 PM

I have also tried with aolserver 4.5.1 (no compiling errors) but the problem continues. And also when it is hanging I check the postgres database if there is some kind of extrange query running, but nothing is happening.

thanks again,

Sergio Izquierdo

8: Re: Moving .LRN from server (response to 1)

Posted by Sergio Izquierdo on 09/10/10 05:34 PM

After a minute of so I have this information on the error log:

[10/Sep/2010:04:27:54][4533.3043560352][-default:2-] Notice: largo 4
[10/Sep/2010:04:28:32][4533.2920283040][-sched:idle0-] Notice: starting
[10/Sep/2010:04:28:32][4533.2920283040][-sched:13-] Notice: dbdrv: opening database 'postgres:localhost::siecadev'
[10/Sep/2010:04:28:32][4533.2920283040][-sched:13-] Notice: Opening siecadev on localhost
[10/Sep/2010:04:28:32][4533.2920283040][-sched:13-] Notice: Ns_PgOpenDb(postgres): Openned connection to localhost::siecadev.
[10/Sep/2010:04:28:35][4533.2915224480][-sched:idle1-] Notice: starting
[10/Sep/2010:04:29:30][4533.2980420512][-default:6-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:29:35][4533.2915224480][-sched:20-] Notice: dbdrv: opening database 'postgres:localhost::siecadev'
[10/Sep/2010:04:29:35][4533.2915224480][-sched:20-] Notice: Opening siecadev on localhost
[10/Sep/2010:04:29:35][4533.2915224480][-sched:20-] Notice: Ns_PgOpenDb(postgres): Openned connection to localhost::siecadev.
[10/Sep/2010:04:29:39][4533.3060513696][-default:1-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:29:39][4533.3065572256][-default:0-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:29:57][4533.3028384672][-default:4-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:30:03][4533.3033443232][-default:7-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:30:08][4533.2985479072][-default:5-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:30:08][4533.2990537632][-default:8-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:30:12][4533.3038501792][-default:3-] Notice: exiting: timeout waiting for connection
[10/Sep/2010:04:31:29][4533.3043560352][-default:2-] Warning: db_exec: longdb 214670501 nsdb0 select dbqd.dotlrn.tcl.module-grades-procs.sieca_modules::get_assessment_grade.get_assessments
[10/Sep/2010:04:31:29][4533.3043560352][-default:2-] Notice: VAAA grade 0.0 min_grade 70 enabled_p 0 ponderation 25 grade_compare 17
[10/Sep/2010:04:31:29][4533.3043560352][-default:2-] Notice: VAAAA approved_p 0 grade 0.0 grade_compare 17

Don't know if it helps.

Sergio

9: Re: Moving .LRN from server - Good News & Bad News (response to 1)

Posted by Sergio Izquierdo on 09/10/10 06:59 PM

I have also checked the Postgress activity and there is no query running that could hang the aplication

siecadev=# select * from pg_stat_activity;
datid | datname | procpid | usesysid | usename | current_query | query_start
-------+----------+---------+----------+----------+---------------+-------------
17143 | siecadev | 4726 | 1 | postgres | |
17143 | siecadev | 4669 | 1 | postgres | |
17143 | siecadev | 4557 | 1 | postgres | |
17143 | siecadev | 4555 | 1 | postgres | |
17143 | siecadev | 4538 | 1 | postgres | |
17143 | siecadev | 4537 | 1 | postgres | |
17143 | siecadev | 4536 | 1 | postgres | |
(7 rows)

siecadev=#

Maybe other people is having the same problem:
http://panoptic.com/aolserver/chat/20080715.html

10: Re: Moving .LRN from server - Good News & Bad News (response to 9)

Posted by Gustaf Neumann on 09/11/10 11:25 AM

not sure, what your problem is: is it slow or does it "crash" where the term "crash" is sometimes used for an uncaught error message in the browser and sometimes for a disappearing server process.

i got the impression from posting #8 that it is slow. even if the number in "db_exec: longdb 214670501 ..." is a time in microseconds, this is quite long. This seems to be some custom query in "sieca_modules" (whatever this is) which is not part of dotlrn.

The message from the referenced chat "timeout waiting for connection" is nothing to be worried about; it simply says that a connection thread stopped because it did not receive anything to do in the specified timeout interval. The sentence about the "db pool" is maybe from a wrongly configured db, but has most likely nothing to do with your problem, since you could connect to your database (... "I can add new courses...").

my guess is that something is substantially wrong with the "sieca_modules", maybe some indices missing, etc.