RE: CRCOLS and VERCOLS and REPLCHECK
Hey All,
I was just experimenting with the replcheck cols to improve sync
performance, and I seem to get a segmentation fault when trying to do a
cdr check replicate.
Nothing relevant seems to show up in the online log on either server.
[informix@r1) ~]$ cdr check replicate --master g_s67 --repl=aqs_packs_67
g_r1;
Apr 23 2010 09:13:20 ------ Table scan for aqs_packs_67 start
--------
Segmentation fault
------------------
We are running on RHEL 4 with informix 11.50.UC5 / (X2)
The X2 patch fixed an sever memory leak we had with ER transmits. It is
only installed on s67 as r1 doesn't send much er data so memory leak
doesn't really affect it.
R1 is and enterprise er server, and s67 is a leaf node WE edition.
[informix@r1 ~]$ onstat -
IBM Informix Dynamic Server Version 11.50.UC5 -- On-Line -- Up
17:48:00 -- 2295016 Kbytes
[sdev@s67 ~]$ onstat -
IBM Informix Dynamic Server Version 11.50.UC5X2 -- On-Line -- Up 20
days 17:42:21 -- 505544 Kbytes
Has anyone else seen this? Could it be due to our X2 patch?
Should also note that packs_new on r1 was an empty table, just created
it to test. I then copied some rows from s67 with a insert into select
* from and tried again...
[informix@r1 ~]$ cdr check replicate --master g_sirrp67
--repl=aqs_packs_67 g_rhea;
Apr 23 2010 10:36:24 ------ Table scan for aqs_packs_67 start
--------
Error in cmpCol
SQL types do not match for pack_id(117) unit_id(103)
command failed -- Source and Target do not have the same data type (202)
ER/DB details:
[informix@r1) ~]$ cdr define repl -C ignore aqs_packs_67 \\
"R aqs@g_r1:sdev.packs_new" "SELECT * FROM packs_new" \\
"P aqs@g_s67:sdev.packs" "SELECT * FROM packs"
[informix@r1) ~]$ cdr modify repl --ignoredel y aqs_packs_67
[informix@r1) ~]$ cdr start repl aqs_packs_67
[informix@r1) ~]$ dbschema -d aqs -t packs_new -ss
DBSCHEMA Schema Utility INFORMIX-SQL Version 11.50.UC5
{ TABLE "sdev".packs_new row size = 102 number of columns = 18 index
size = 51 }
create table "sdev".packs_new
(
pack_id serial8 not null ,
unit_id integer not null ,
item_dt datetime year to fraction(3),
pack_number integer,
check_weight float not null ,
tare_weight float,
head_target float,
head_number integer,
po_number char(12),
productsource integer,
aqs_control smallint
default 0,
valid_pack smallint
default 0,
seal_check smallint
default 0,
metal_detected smallint
default 0,
underweight smallint
default 0,
overweight smallint
default 0,
print_check smallint
default 0
) in aqs extent size 25000000 next size 409600 lock mode page;
alter table "sdev".packs_new add crcols;
alter table "sdev".packs_new add replcheck;
revoke all on "sdev".packs_new from "public" as "sdev";
create unique index "sdev".packs_er_idx on "sdev".packs_new (pack_id,
ifx_replcheck) using btree in aqs;
create unique index "sdev".packs_idx on "sdev".packs_new (unit_id,
item_dt) using btree in aqs;
alter table "sdev".packs_new add constraint primary key (unit_id,
item_dt) constraint "sdev".packs_pk ;
alter table "sdev".packs_new add constraint (foreign key (unit_id)
references "sdev".units constraint "sdev".pack_new_units_fk);
[informix@s67 ~]$ dbschema -d aqs -t packs -ss
DBSCHEMA Schema Utility INFORMIX-SQL Version 11.50.UC5X2
{ TABLE "sdev".packs row size = 102 number of columns = 18 index size =
79 }
create table "sdev".packs
(
pack_id serial8 not null ,
unit_id integer not null ,
item_dt datetime year to fraction(3),
pack_number integer,
check_weight float not null ,
tare_weight float,
head_target float,
head_number integer,
po_number char(12),
productsource integer,
aqs_control smallint
default 0,
valid_pack smallint
default 0,
seal_check smallint
default 0,
metal_detected smallint
default 0,
underweight smallint
default 0,
overweight smallint
default 0,
print_check smallint
default 0
) extent size 40960 next size 40960 lock mode row;
alter table "sdev".packs add crcols;
alter table "sdev".packs add replcheck;
revoke all on "sdev".packs from "public" as "sdev";
create index "sdev".item_dt_idx on "sdev".packs (item_dt) using
btree in data01;
create unique index "sdev".pack_id_idx on "sdev".packs (pack_id)
using btree in data01;
create unique index "sdev".packs_er_idx on "sdev".packs (pack_id,
ifx_replcheck) using btree in data01;
create index "sdev".po_idx on "sdev".packs (po_number) using btree
in data01;
alter table "sdev".packs add constraint (foreign key (unit_id)
references "sdev".units constraint "sdev".pack_units_fk);
From: informix-list-bounces@iiug.org
[mailto:informix-list-bounces@iiug.org] On Behalf Of Art Kagel
Sent: Thursday, April 22, 2010 5:25 AM
To: Laurie Gustin
Cc: informix-list@iiug.org
Subject: Re: CRCOLS and VERCOLS and REPLCHECK
Why choose, you can have all three:
> create table has_it_all(one serial) with crcols with replcheck withvercols;
Table created.
ER needs CRCOLs for normal processing and conflict resolution. It can
use the replcheck columns to speed synchronizing servers. Any and all
can be used by application programs to implement optimistic locking
protocols better and more cheaply.
Art
Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
IIUG Board of Directors (art@iiug.org)
See you at the 2010 IIUG Informix Conference
April 25-28, 2010
Overland Park (Kansas City), KS
www.iiug.org/conf
Disclaimer: Pl