Forum OpenACS Q&A: 3.2.5 with 4.6 on same Linux/Postgres install -- delurk

Can they both co-exist on the same machine? I believe OpenACS 4.6 requires a newer AOLserver and I'm wondering if it is possible to keep both AOLservers running without messing things up too much? Can I just install the new one in a new directory?

On the other hand, I don't mind installing the new server on top of the old one if it's the most stable way to go. Is it?

Thanks to all for the excellent OpenACS...

Yes, both can coexist without a problem. Just make sure of using a different database for each ;)

Cheers

/B

Thanks a bunch

which produced a 30 MB tarball.

I installed RH 7.3 and PostgreSQL 7.2.4 rpm version.

I created a pristine database:

createdb -T template0 newdb

Then:

pg_restore -d mydb.tar newdb

This produced a lot of notices, until:

pg_restore: NOTICE:  CREATE TABLE / PRIMARY KEY will create implicit index 'event_info_pkey' for table 'event_info'
pg_restore: ERROR:  copy: line 389, Cannot insert a duplicate key into unique index user_group_map_pkey
pg_restore: lost synchronization with server, resetting connection
pg_restore: [archiver (db)] error returned by PQendcopy

And that was it.

psql -V output:

psql (PostgreSQL) 7.2.4
contains support for: readline, history, multibyte
Portions Copyright (c) 1996-2001, PostgreSQL Global Development Group
Portions Copyright (c) 1996, Regents of the University of California
Read the file COPYRIGHT or use the command \copyright to see the
usage and distribution terms.

A smaller OpenACS 3.2.5 database does load without errors and is OK.

An OpenACS4 database did not load at all:

$ pg_restore -d oacs oacs46-12mrt.tar
pg_restore: [archiver (db)] could not execute query: ERROR:  Function 'tree_ancestor_keys(varbit, int4)' does not exist
    Unable to identify a function that satisfies the given argument types
    You may need to add explicit typecasts

Anybody got a clue? Thanks!

P.S. Maybe I should post this at the PostgreSQL forum as well...

[14/Mar/2003:23:19:18][5118.4101][-conn0-] Error: dbinit: error(localhost::blog,ERROR: value too long for type character(1) ): 'select sec_rotate_last_visit(8780, 1047680358)' [14/Mar/2003:23:19:18][5118.4101][-conn0-] Error: Filter sec_read_security_info returned error #1: Database operation "1row" failed (exception NSDB, "Query was not a statement returning rows.") [14/Mar/2003:23:19:18][5118.4101][-conn0-] Error: tclop: invalid return code from filter proc 'Critical filter sec_read_security_info failed.': must be filter_ok, filter_return, or filter_break

Very annoying.... By the way, I do not expect any of you to have a solution for this... I guess we just have to use the old databases on a different install.

Actually, there is a quite simple fix for this problem. In the function sec_rotate_last_visit in security.sql, just cast the variable vtime as a varchar, not a char, thus:

create function sec_rotate_last_visit(integer, integer)
returns integer as '
DECLARE
	v_browser_id alias for $1;
	v_time alias for $2;
BEGIN
    lock table sec_browser_properties;
    delete from sec_browser_properties
        where browser_id = v_browser_id and module = ''acs'' and property_name = ''second_to_last_visit'';
    update sec_browser_properties
        set property_name = ''second_to_last_visit''
        where module = ''acs'' and property_name = ''last_visit'' and browser_id = v_browser_id;
    insert into sec_browser_properties(browser_id, module, property_name, property_value, secure_p)
        values(v_browser_id, ''acs'', ''last_visit'', v_time::varchar, ''f'');

    return 1;
end;
' language 'plpgsql';

Drop the function using PSQL and then read the security.sql file into the database to reload the function. I think all will be happiness and light then.

This actually has been previously discussed at length. See https://openacs.org/forums/message-view?message_id=29185 for one thread...