By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
438,852 Members | 2,198 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 438,852 IT Pros & Developers. It's quick & easy.

New eRServer patch/fork release

P: n/a

A new build of my modifications to eRServer is available at
http://grayling.ravensfield.com. Changes/fixes:

* Fixed existing column as unique problem
I had a bug in the replication threads that manifested itself if you
used existing columns as the unique column. Caused everything to stop.

* Added RMI signaling structure
There is now an RMI-based signaling structure to the whole thing.
Currently, it only supports shutdown, but will eventually support
reseting the
various column caches and table maps so you can modify table schemas
without having to restart erserver. There is no security at all on it
yet. I
probably ought to require the master password for remote operations.

* Added org.psotgresql.ers.ERSCtl
Signaling app.

* Modified ers_ctl to use the shutdown signal
Cleaner way to shut the thing down. Using the current ers_ctl will
still work fine (except when it screws up on its own accord).

* Added org.postgresql.ers.ERSExecSQL
This will execute an arbitrary SQL statement on the master and slaves
(default is slaves first). When I get the signaling down proper, this
will allow
you to add columns etc. on the fly.

* Added by-table unique column setting while adding multiple tables
from a file or command line
So you can do the following things:

java org.postgresql.ers.ERSAddTable table1 table2,col1

which will add table1 using '_ers_uniq' and table2 using 'col1', or:

java org.postgresql.ers.ERSAddTable -u _uniq table1 table2,col1 table3

which will add table1 and table3 with '_uniq' and table2 with col1.

Questions, comments, bugs, flames, etc. let me know.

--------------------

Andrew Rawnsley
President
The Ravensfield Digital Resource Group, Ltd.
(740) 587-0114
www.ravensfield.com
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

Nov 12 '05 #1
Share this Question
Share on Google+
2 Replies


P: n/a
On Thu, Nov 06, 2003 at 10:21:32AM -0500, Andrew Rawnsley wrote:

A new build of my modifications to eRServer is available at
http://grayling.ravensfield.com. Changes/fixes:
Could you add these patches to the gborg site? That way people who
go there to test it out won't overlook it.
* Added org.postgresql.ers.ERSExecSQL
This will execute an arbitrary SQL statement on the master and slaves
(default is slaves first). When I get the signaling down proper, this
will allow
you to add columns etc. on the fly.


You'll need to change the engine. The engine currently only looks
for columns at the time of start-up. After that, it never looks
again, so the columns will be added, but not replicated.
A
----
Andrew Sullivan 204-4141 Yonge Street
Afilias Canada Toronto, Ontario Canada
<an****@libertyrms.info> M2P 2A8
+1 416 646 3304 x110
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

Nov 12 '05 #2

P: n/a

On Nov 6, 2003, at 10:48 AM, Andrew Sullivan wrote:
On Thu, Nov 06, 2003 at 10:21:32AM -0500, Andrew Rawnsley wrote:

A new build of my modifications to eRServer is available at
http://grayling.ravensfield.com. Changes/fixes:
Could you add these patches to the gborg site? That way people who
go there to test it out won't overlook it.


Sure. Should have thought to do that.
* Added org.postgresql.ers.ERSExecSQL
This will execute an arbitrary SQL statement on the master and slaves
(default is slaves first). When I get the signaling down proper, this
will allow
you to add columns etc. on the fly.


You'll need to change the engine. The engine currently only looks
for columns at the time of start-up. After that, it never looks
again, so the columns will be added, but not replicated.


Yeah, actually reseting the column cache in the slave processor and the
table map in the master is easy. Getting to those objects
through the rather obtuse layering of classes from where I put the
signaling point is the trickier part.

A
----
Andrew Sullivan 204-4141 Yonge Street
Afilias Canada Toronto, Ontario Canada
<an****@libertyrms.info> M2P 2A8
+1 416 646 3304 x110

_______________________________________________
Erserver-general mailing list
Er**************@gborg.postgresql.org
http://gborg.postgresql.org/mailman/...server-general

--------------------

Andrew Rawnsley
President
The Ravensfield Digital Resource Group, Ltd.
(740) 587-0114
www.ravensfield.com
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match

Nov 12 '05 #3

This discussion thread is closed

Replies have been disabled for this discussion.