Skip site navigation (1) Skip section navigation (2)

Re: Updated patch: Add support for execution of jobs on a remote database

From: Ashesh Vashi <ashesh(dot)vashi(at)enterprisedb(dot)com>
To: Dave Page <dpage(at)pgadmin(dot)org>
Cc: pgadmin-hackers <pgadmin-hackers(at)postgresql(dot)org>
Subject: Re: Updated patch: Add support for execution of jobs on a remote database
Date: 2008-12-26 09:57:48
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgadmin-hackers
Hi All,

Please find the updated patches for the same as per the instructions 
from Dave Page.
>> - Add support for execution of jobs on a remote database in pgAgent.
>>   (primarily for hot standby support). This essentially means allowing a
>>   full connection string to be specified in place of the current database
>>   name option.
> <snip comprehensive description>
> Hi Ashesh,
> I haven't tested fully at this stage, just eyeball reviewed the code,
> made sure it builds (on Windows) and checked out the UI. First
> impression is that it's well thought out and nicely implemented. I do
> have a few comments though - mostly minor:
> - The pgTable::HasColumn() static function seems like an unusual fit
> in that class. Yes, it's a table object, but that class is primarily
> concerned with representing a table in the treeview. It seems to me
> that a non-static pgConn::TableHasColumn() function might make more
> sense, and would probably look cleaner at the call sites.
> - The database icon in dlgSelectDatabase seems to be missing (I just
> get a blank space where it should be).
> - The height of the connection string textbox is inconsistent with all
> other textboxes. It should be sized as per the other one line text
> boxes, and the button altered to match it.
> - Any new headers should be added to include/precomp.h (we all forget
> to do that!)
> - In pgAgent, why is connInfo declared as a struct and not a class?
> - pgAdmin seems to be able to handle connecting to an old-style schema
> (which is good), but I didn't see any code in pgAgent to do similar. I
> would suggest we do something like the following, which should allow
> for future changes:
>   * Add an SQL function pgagent_schema_version() which returns an int
> (with a value of 3 for this version - we'll bump the package to
> 3.0.0).
Introduced a new function pgagent_schema_version() in the pgagent.sql
Also, written into the pgagent_update.sql for unreadability.
>   * Add a check to pgAgent in (MainLoop()) - upon setup of the primary
> connection, if pgagent_schema_version() does not exist, then exit with
> an error asking the user to upgrade the schema.
Checking for pgagent.pgagent_schema_version exists or not and also check if
return value of this function matches with the major version of pgagent.
>   * If it does exist, check the value it returns against MAJOR_VERSION
> - a pre-processor macro we can set in cmake from
> CPACK_PACKAGE_VERSION_MAJOR - if it doesn't match, exit with an error
> telling the user there is a schema version mismatch.
Made changes in CMakeList.txt for adding definition of 

BTW, changed the version of pgagent to 3.0.1 in the CMakeLists.txt

Ashesh Vashi

EnterpriseDB INDIA:

Attachment: pgadmin_agent_remotedb_v2.patch
Description: text/x-patch (37.7 KB)
Attachment: pgagent_remotedb_v2.patch
Description: text/x-patch (22.4 KB)

In response to


pgadmin-hackers by date

Next:From: Ashesh VashiDate: 2008-12-26 10:03:28
Subject: Re: Updated patch: Add support for execution of jobs on a remote database
Previous:From: svnDate: 2008-12-19 11:12:54
Subject: SVN Commit by dpage: r7524 - trunk/www/download

Privacy Policy | About PostgreSQL
Copyright © 1996-2017 The PostgreSQL Global Development Group