Problems with Data Replication
Posted in 2000
Topics: High Availability & Replication, Storage & Space Management, Server Administration, Networking & sqlhosts Configuration, Versions, Editions & End-of-Life
I have problem with data replication between 2 database servers.
I have two computers ( alfa and atlas) : Both have IDS 7.30 FC7.
Each host permit connections for shared memory and for tcp with sockets.
(onipcshm and onsoctcp) ; the definitions on sqlhosts file, host file,
host_equiv,.rhost
and services are corrected ( I believe ) because I can connect to the hosts
in dbaccess
utility.
I've defined two entrys more on each sqlhosts file for data replication .
This entrys are
"groups". Each group belongs to the connections for one host:
sqlhost files: (are equal in 2 hosts).
dbservername nettype host servicename options
g_atlas group - - e=atlas_tcp,i=1
atlas_sh onipcshm atlas atlas_sh
atlas_tcp onsoctcp atlas atlas_tcp
g_alfa group - - e=alfa_tcp,i=2
alfa_sh onipcshm alfa alfa_sh
alfa_tcp onsoctcp alfa alfa_tcp
Then I've tried to configure data replication between 2 groups of servers.
Really I need only unidirettional replication (primary to target): for alfa
to atlas; but I don't know how do this.
I've use the CLU to try configure the replication.
I've defined the two servers ( or only I need define the primary server and
no the target server?)
with the CLU in alfa host:
"cdr def server --init
g_alfa --ats=/usr/informix/cdr_alfa/ats_alfa -R
/usr/informix/cdr_alfa/ris_alfa" -> OK
The IDS create the syscdr database and the syscdr* views on sysmaster
database on host alfa
"cdr def server --connect g_atlas --init
g_atlas --ats=/usr/informix/cdr_atlas/ats_atlas
-R /usr/informix/cdr_atlas/ris_atlas " -> OK
The IDS create the syscdr database and the syscdr* views on
sysmaster database on host atlas
Then, I define the replication with:
" cdr define repl --connect=g_alfa -C ignore repl1
"asemas@g_beta:ceinsa.arquitecto"
"select * from arquitecto" -> OK
Really I want to define the unidirettional replication from alfa to atlas.
Finally I start the replication with "cdr start repl --connect=g_alfa repl1"
If I execute "cdr list server":
SERVER ID STATE STATUS CONNECTION CHANGED
----------------------------------------------------------------------------
-
g_alfa 2 Active Local
If I execute "cdr list server -c g_atlas":
SERVER ID STATE STATUS CONNECTION CHANGED
----------------------------------------------------------------------------
-
g_atlas 1 Active
If I execute "cdr list repl":
REPLICATE STATE CONFLICT FREQUENCY OPTIONS
----------------------------------------------------------------------------
repl1 ACTIVE ignore immediate row
All is Ok; but when I make an insert into the arquitecto table on server
alfa, no
replication occurrs to server atlas and the online.log on server alfa is :
18:11:14 CDR Queuer: Is suspended failed gc_queryGroup servers returned 0!!
for group/repl: 131073, commit time: 959789473, send time: 959789473, ret
code: 0. 18:11:14
CDR Grouper Evaluator could not give buffer to Queuer.
What is THIS?. May be a problem with the dbspaces????
Perhaps I need sync the two computers ????
When I try to sync the two servers , I have a message that says: alfa and
atlas differs -400 seconds.
Thank You and sorry for my poor english.
I suppose we are talking Enterprise Replication here & not Continuous
Data Replication, if you are using the cdr command line utility &
talking about primary to target strategies, etc.?
Yes, use the rdate command to sync the servers as a first step before
even defining servers - strange errors result otherwise!!! Also, I
would recommend upgrading to 7.31 if possible, as we found ER much more
stable in 7.31 than on 7.30.
Lastly, nothing wrong with your english - very good for a "non-english-
first-language" person.
Eben
In article <8q7v0q$8oq$1@unbe.sarenet.es>,
"Luisma Sanchez" <luismas@asemas.es> wrote:
> I have problem with data replication between 2 database servers.
>
> I have two computers ( alfa and atlas) : Both have IDS 7.30 FC7.
>
> Each host permit connections for shared memory and for tcp with
sockets.
> (onipcshm and onsoctcp) ; the definitions on sqlhosts file, host file,
> host_equiv,.rhost
> and services are corrected ( I believe ) because I can connect to the
hosts
> in dbaccess
> utility.
>
> I've defined two entrys more on each sqlhosts file for data
replication .
> This entrys are
> "groups". Each group belongs to the connections for one host:
>
> sqlhost files: (are equal in 2 hosts).
>
> dbservername nettype host servicename options
>
> g_atlas group - - e=atlas_tcp,i=1
> atlas_sh onipcshm atlas atlas_sh
> atlas_tcp onsoctcp atlas atlas_tcp
> g_alfa group - - e=alfa_tcp,i=2
> alfa_sh onipcshm alfa alfa_sh
> alfa_tcp onsoctcp alfa alfa_tcp>
> Then I've tried to configure data replication between 2 groups of
servers.
> Really I need only unidirettional replication (primary to target):
for alfa
> to atlas; but I don't know how do this.
> I've use the CLU to try configure the replication.
>
> I've defined the two servers ( or only I need define the primary
server and
> no the target server?)
> with the CLU in alfa host:
> "cdr def server --init
> g_alfa --ats=/usr/informix/cdr_alfa/ats_alfa -R
> /usr/informix/cdr_alfa/ris_alfa" -> OK
>
> The IDS create the syscdr database and the syscdr* views on sysmaster
> database on host alfa
>
> "cdr def server --connect g_atlas --init
> g_atlas --ats=/usr/informix/cdr_atlas/ats_atlas
>
> -R /usr/informix/cdr_atlas/ris_atlas " -> OK
>
> The IDS create the syscdr database and the syscdr* views on
> sysmaster database on host atlas
>
> Then, I define the replication with:
> " cdr define repl --connect=g_alfa -C ignore repl1
> "asemas@g_beta:ceinsa.arquitecto"
> "select * from arquitecto" -> OK
>
> Really I want to define the unidirettional replication from alfa to
atlas.
>
> Finally I start the replication with "cdr start repl --connect=g_alfa
repl1"
>
> If I execute "cdr list server":
>
> SERVER ID STATE STATUS CONNECTION CHANGED
> ----------------------------------------------------------------------
------
> -
> g_alfa 2 Active Local
>
> If I execute "cdr list server -c g_atlas":
>
> SERVER ID STATE STATUS CONNECTION CHANGED
> ----------------------------------------------------------------------
------
> -
> g_atlas 1 Active
>
> If I execute "cdr list repl":
>
> REPLICATE STATE CONFLICT FREQUENCY OPTIONS
> ----------------------------------------------------------------------
------
> repl1 ACTIVE ignore immediate row
>
> All is Ok; but when I make an insert into the arquitecto table on
server
> alfa, no
> replication occurrs to server atlas and the online.log on server alfa
is :
>
> 18:11:14 CDR Queuer: Is suspended failed gc_queryGroup serversreturned 0
> !!
> for group/repl: 131073, commit time: 959789473, send time:
959789473, ret
> code: 0.
18:11:14
> CDR Grouper Evaluator could not give buffer to Queuer.
>
> What is THIS?. May be a problem with the dbspaces????
> Perhaps I need sync the two computers ????
>
> When I try to sync the two servers , I have a message that says: alfa
and
> atlas differs -400 seconds.
>
> Thank You and sorry for my poor english.
>
>
Sent via Deja.com http://www.deja.com/
Before you buy.