Rac loadbalancing/failover - services

I have seen some DBA’s creating 2 services for every rac database for example in a 2 node rac they would have.
Svcprd1 prefered node: node 1 available node:node2
Svcprd2 prefered node: node2 available node:node1
What’s the difference between above and having one service with both nodes as preferred nodes?
Does the first method provide better HA ? or extra flexibility ?

I'm not sure what the two services are used for, but if one is for oltp and the other is for batch, then the services separate the two activities so they don't consume resources ram/cpu on the same none. The alternate node serves as a failover if the primary node goes down.

Similar Messages

Oracle11g r2 Grid/RAC VIP failover instead of SCAN VIP failover

Dear Experts and Gurus
Our Platform: 2-Node ORACLE11G r2 RAC/GRID 11.2.0.1.0
ReadHat Enterprise Linux5.3 64 bit
We have not available the DNS Server for used to SCAN feature of Oracle11g r2 GRID/RAC.
we have successfully deployed the the setup using scan-vip in /etc/host in our production site.
we want to used the Oracle11g r2 Grid/RAC as Oracle10g r2 RAC/Oracle11g r1 RAC(VIP Failover)
plz find the default configurations of my setup.
cat /etc/hosts
#public
xxx.xxx.0.1 xyz-ch-aaadb-01
xxx.xxx.0.2 xyzl-ch-aaadb-02
#Virtual
xxx.xxx.0.3 xyz-ch-aaadb-01-vip
xxx.xxx.0.4 xyz-ch-aaadb-02-vip
#Private
10.10.0.1 xyz-ch-aaadb-01-priv
10.10.0.2 xyz-ch-aaadb-01-priv
#Scan
xxx.xxx.0.5 rac-scan
cat listener.ora
listener.ora in both the RAC nodes
LISTENER=(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=IPC)(KEY=LISTENER)))) # line added by Agent
LISTENER_SCAN1=(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=IPC)(KEY=LISTENER_SCAN1)))) # line added by Agent
ENABLE_GLOBAL_DYNAMIC_ENDPOINT_LISTENER_SCAN1=ON # line added by Agent
ENABLE_GLOBAL_DYNAMIC_ENDPOINT_LISTENER=ON # line added by Agent
cat tnsnames.ora.
AAADB =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = rac-scan)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = aaadb)
aaadb2 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = xyz-ch-aaadb-02-vip)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = aaadb)
(INSTANCE_NAME = aaadb2)
aaadb1 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = xyz-ch-aaadb-01-vip)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = aaadb)
(INSTANCE_NAME = aaadb1)
listener parameters
RAC-NODE1
SQL> show parameter listener
NAME TYPE VALUE
listener_networks string
local_listener string (DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=xyz-ch-aaadb-01-vip)(PORT=1521))))
remote_listener string rac-scan:1521
RAC-NODE2
SQL> show parameter listener
NAME TYPE VALUE
listener_networks string
local_listener string (DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=xyz-ch-aaadb-02-vip)(PORT=1521))))
remote_listener string rac-scan:1521
listener status
RAC Node-1
[oracle@aaarac1 ~]$ lsnrctl status
LSNRCTL for Linux: Version 11.2.0.1.0 - Production on 30-AUG-2011 23:43:50
Copyright (c) 1991, 2009, Oracle. All rights reserved.
Connecting to (ADDRESS=(PROTOCOL=tcp)(HOST=)(PORT=1521))
STATUS of the LISTENER
Alias LISTENER
Version TNSLSNR for Linux: Version 11.2.0.1.0 - Production
Start Date 30-AUG-2011 22:31:34
Uptime 0 days 1 hr. 12 min. 15 sec
Trace Level off
Security ON: Local OS Authentication
SNMP OFF
Listener Parameter File /u01/app/11.2.0/grid/network/admin/listener.ora
Listener Log File /u01/app/oracle/diag/tnslsnr/aaarac1/listener/alert/log.xml
Listening Endpoints Summary...
(DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=LISTENER)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=xyz-ch-aaadb-01-vip)(PORT=1521)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=xxx.xxx.0.1)(PORT=1521)))
Services Summary...
Service "+ASM" has 1 instance(s).
Instance "+ASM1", status READY, has 1 handler(s) for this service...
Service "aaadb" has 1 instance(s).
Instance "aaadb1", status READY, has 1 handler(s) for this service...
Service "aaadbXDB" has 1 instance(s).
Instance "aaadb1", status READY, has 1 handler(s) for this service...
The command completed successfully
RAC Node-2
[oracle@aaarac2 ~]$ lsnrctl status
LSNRCTL for Linux: Version 11.2.0.1.0 - Production on 30-AUG-2011 23:44:27
Copyright (c) 1991, 2009, Oracle. All rights reserved.
Connecting to (ADDRESS=(PROTOCOL=tcp)(HOST=)(PORT=1521))
STATUS of the LISTENER
Alias LISTENER
Version TNSLSNR for Linux: Version 11.2.0.1.0 - Production
Start Date 30-AUG-2011 22:08:45
Uptime 0 days 1 hr. 35 min. 42 sec
Trace Level off
Security ON: Local OS Authentication
SNMP OFF
Listener Parameter File /u01/app/11.2.0/grid/network/admin/listener.ora
Listener Log File /u01/app/oracle/diag/tnslsnr/aaarac2/listener/alert/log.xml
Listening Endpoints Summary...
(DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=LISTENER)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=xyz-ch-aaadb-02-vip)(PORT=1521)))
(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=xxx.xxx.0.2)(PORT=1521)))
Services Summary...
Service "+ASM" has 1 instance(s).
Instance "+ASM2", status READY, has 1 handler(s) for this service...
Service "aaadb" has 1 instance(s).
Instance "aaadb2", status READY, has 1 handler(s) for this service...
Service "aaadbXDB" has 1 instance(s).
Instance "aaadb2", status READY, has 1 handler(s) for this service...
The command completed successfully
plz suggest the provide the step to configure the listener.ora and tnsnames.ora for use the Oracle11g r2 Grid/RAC to use as
VIP failover instead of SCAN-VIP failover.
Regards
Hitgon
Edited by: hitgon on Aug 31, 2011 12:14 AM

hitgon wrote:
Dear Experts and Gurus
plz suggest the provide the step to configure the listener.ora and tnsnames.ora for use the Oracle11g r2 Grid/RAC to use as
VIP failover instead of SCAN-VIP failover.
Regards
Hitgon
Hi,
Have a read http://download.oracle.com/docs/cd/E11882_01/network.112/e10836/advcfg.htm#NETAG348
Hope it helps
CHeers

Running Distributed Transactions in RAC without DTP service.

We just switched over to our standby database this morning which we had setup to facilitate our ASM migration. The switchover went smooth and the standby is up and running now as my primary.
This is a 2 node rac on linux itanium 10.2.0.3.
I got the following message in the alert log on both instances and I have never seen this message before in my original clustrer?
Running Distributed Transactions in RAC without DTP service.
Is this a reason ofr concern?
Thanks, any feedback is very much appreciated.

Duplicate thread..
Re: Running Distributed Transactions in RAC without DTP service

Is SQL Server Engine failover service spported for Orchestrator databases?

Hi there,
I have the following question:
Is SQL Server Engine failover service spported for Orchestrator 2012 R2 databases?
Thanx in advance.
Regards,
Chris

Hi,
yes, the issue that Orchestrator does not reconnect when the SQL server is restarted is has been fixed in System Center 2012 Orchestrator Service Pack 1 (http://support.microsoft.com/kb/2768366)
Orchestrator requires "only "the basic SQL Server features found in the Database Engine Service. No additional features are required.
Regards,
Stefan
www.sc-orchestrator.eu ,
Blog sc-orchestrator.eu

Independent Zone Paths for Failover Services

We wish to use Sun Cluster 3.2 to operate failover zones across 3 or more servers. It would seem that for failover services, zone paths must exist on shared storage and all associated failover zones must use that exact zone filesystem. However for load-balancing, it appears that multiple independent zonepaths are allowed.
My question is can failover zones be deployed WITHOUT using a single zone path filesystem. It seems to me that it complicates patching and maintenance to the zone filesystem if it is always in use. Not to mention, a corruption in that zone path's files would be forced on the entire set of zones using that path.
I'm new to Sun Cluster and am skittish about deploying a large fiber channel infrastructure if it is not really necessary. Thanks for any corrections / advice on the above.

Now I understand what you want to achieve. Unfortunately wording is not yet precise with all these new technologies.
What you need is not the HA Container agent, and what you want is not failover zones. But you want to have local zones, and let HA services failover between local zones on different nodes. Nothing easier than that. Just use the "normal" procedures to configure an agent. Now, instead of using the nodenames for the nodelist of the resource group, give it a list of the zone names that you want this service to run in. Example could be:
clrg create -p nodelist=host-A:zone-FOO,host-B:zone-FOO,host-C:zoneFOO. Now you can switch the HA service between the zone-FOO(s) on the 3 nodes without switching the zones themselves. And you can patch them without taking the service offline. Great.
If you want to make the data that your HA service uses highly available, you still would need some kind of shared filesystem. How could the HA service, switched onto another node:zone would get access to it, if not via a filesystem that was also switched over?
There is still value to the HA COntainer agent: if, e.g. you want to use the blackbox approach and do not care much about the potential outages of the zones during patch cycles. And there is more.
Hope that helps
Hartmut

Handling Dyn DNS Active/Failover service while also hosting internal DNS

I want to try and take advantage of using Dyn DNS failover service for our websites, to where it will detect when our primary public IP address to webserver goes down and will automatically update the dns record to use our 2ndary public IP address in our
failover site. The only trouble i'm running into is we also host the domain name internally as well which we normally point to the private of the web server in our primary site. is there a way to just have the request for that one specific A record
to go to an external dns server to get resolved while the remaining records can be resolved internally by that server? tried round robin with the 2 ip's addresses but does not work as i need it to. Thanks any help is appreciated.
Michael Duhon

Hi,
According to your description, my understanding is that you want the customer to access the website by another public IP when the current public IP down: request for specific A record to go to an external DNS server to get resolved while the remaining records
can be resolved internally by that server.
Usually we can use DNS conditional forwarder to redirect the query for a specific domain name to another DNS server, but we can’t specific a DNS server for an A record query.
Or, you may try to manually add the record in local hosts file.
Best Regards,
Eve Wang
Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Support, contact [email protected]

ALBPM 5.7 BPM process admin issues with Oracle 11g RAC as directory service

Hi,
Setup ALBPM 5.7 enterprise for weblogic.
BPM Engine deployed on weblogic cluster.
BPM process admin (web console) deployed on tomcat.
BPM process admin uses WEB-INF/directory.properties for directory configuration.
things were fine with oracle SID,
directory.default.url=oracle://<db_host>:1521/schema=diruser,sid=BPMDB { with this, login to process admin was working fine}
But clueless, on how to configure RAC as the directory service in the above directory.properties file. Since this is a old version, no directory.xml in the installation.
Connection string:
(DESCRIPTION =
(LOAD_BALANCE = on)
(FAILOVER = on)
(ADDRESS = (PROTOCOL = TCP)(HOST = DB_HOST1)(PORT=1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = DB_HOST2)(PORT=1521))
(CONNECT_DATA =
(SERVICE_NAME = BPMDB)
(FAILOVER_MODE =
(TYPE = SELECT)
(METHOD = BASIC)
Please help how to use service name (above connection string) in directory.properties, to be able to login to process admin successfully.
Thanks in advance,
Swamy

Using admin center, able to get the string to be used in directory.properties.
FYI, the string looks like below,
directory.default.url=oracle://customURL:0/schema=bpmdirectory,customURL=jdbc:oracle:thin:@(DESCRIPTION = (LOAD_BALANCE = on)(FAILOVER = on)(ADDRESS = (PROTOCOL = TCP)(HOST = DB_HOST1)(PORT=1521))(ADDRESS = (PROTOCOL = TCP)(HOST = DB_HOST2)(PORT=1521)) (CONNECT_DATA = (SERVICE_NAME = BPMDB.DOMAIN.COM) (FAILOVER_MODE = (TYPE = SELECT) (METHOD = BASIC))))

RAC DB failover from one Datacenter to Standby site

Hi,
I have 2 node 10gR2 RAC cluster running on Linux with ASM and E-business suite is configured with this 10gR2 RAC DB. Would like to setup standby site with single instance and not go with RAC for now until we procure additional hardware. Also no plans in setting up DG as solution. Does anyone has similar setup? During failover we have the capability to migrate the VIP's (names) to the failed DC but the IP's will change since there is no L2 extension b/w DC's to retain the same IP. Synchronous replication is setup at disk level using EMC SRDF
Question:
1) Will i be able to bring up the RAC DB in single instance node at second DC using non-rac code tree? IF so can anyone share whitepaper/docs on how to bring up RAC db as non-RAC environment. Logically i think it should work if i have right spfile with no RAC specific parameters.. Would like to hear from experts.
2) Wanted to minimize the configuration changes at the E-business suite level since we plan to bring the VIP's over to failed DC.
3) Does the crs registry manages information about VIP name to IP address? since my IP address will change when i migrate the VIP name from one DC to other DC. Will it be relevant at all since we plan to bring up the db in single instance mode at the failed over DC?
Any help is appreciated
Thanks

1) When you say you want to have it in a non-RAC mode. what do you mean, your current DR is a RAC however during disaster you want it to be in a non-RAC mode. Yes it would be possible, you can turn off the CLUSTER_DATABASE parameter to FALSE and bring the database as a single instance database. with this option you still will have the RAC features such as dictionaries etc however not used. the other option of completely turning of the RAC feature and getting it back to a single instance configuration may require additional steps.
which ever method you pick, you should consider the workload, will this single instance be able to handle the entire workload that originally existed on a RAC configuration.
2) VIPs dont failover outside the RAC cluster. the method normally followed is to have a database service configured on the primary and as part of a database startup trigger ensure that the database service is enabled on the DC site. There is a paper on OTN on how to do this.
3) Yes, the VIP information is stored in the OCR file.

Oracle Forms 10g runtime handling during RAC node failover.

Hi,
Forms version 10g R2 (10.1.2.0.2)
Oracle DB version 10g R2 RAC with 3 nodes.
If the RAC DB node that the user is connected to goes down, the user gets FRM-40733 and ORA-03114 error messages and the client forms application gets locked down/ goes in a loop with the error messages. The user has to close the browser to get out of the loop. I understand that this is the expected behaviour, but I'm wondering whether we can trap the error ORA-03114 and fire the "key-exit" trigger to get out of the application.
Have any one implemented a clean way to exit the Forms application when the RAC DB node goes down..?
I'm looking for some suggestions or an elegant way to handle the above failure.
Thank you in advance.
Sudhakar

Glen,
I haven't solve this one yet.I have been playing around with the following:
In my environment, I am still using 6i (not web) forms/reports.
My clients are XP, NT, 2000.
I have the forms/report runtime installed on their PCs.
Their TNSNAMES.ORA will be pointing to PRIMARY ( PDB)
If a SWITCHOVER or FAILOVER happens to the physical standby (SDB), I want a trigger to kick a batch file that will manipulate the TNSNAME.ORA on each clients stations.
On the standby
CREATE OR REPLACE TRIGGER change_tns
AFTER DB_ROLE_CHANGE ON DATABASE
DECLARE
role VARCHAR2(30);
dbname varchar2(100);
BEGIN
SELECT
DATABASE_ROLE,
DB_UNIQUE_NAME
INTO
role,
dbname
FROM
V$DATABASE;
IF role = 'PRIMARY' and dbname='SDB' THEN
dbms_scheduler.create_job(
job_name=>'move_sqlnet',
job_type=>'executable',
job_action=>'c:\temp\movetns.cmd',
enabled=TRUE
ELSE
-- if the standby >was< PRIMARY,
-- but the primary comes BACK on line,
-- need to reverse the step above.
END IF;
END;
As for the movetns.cmd
something like
rem -- attach to the workstation,
net use m: \\station name\share name
rem -- stdb_tnsname.ora would be pointing to STANDBY
copy stdb_tnsname.ora m:\orant\net80\tnsname.ora
net use m: /delete
rem -- need to do that for all workstations..
As you can see, there could be lots of problems with this procedure.
Client doesn't know about the failover, starts a RE-BOOT on the pc, therefore, the new tnsnames.ora will not get to client.. what to do for that client? Do i re-run the batch ...every hour?
tell me if you come up with an answer..
p-

Oracle RAC installation failover

Hi,
I have an Oracle RAC installation with 2 nodes with the data stored on a shared OCFS partition. I had a client test the connection using jdbc string for RAC failover. I tried shutting down one of the nodes on the RAC installation and the client could not connect to the oracle cluster database for the next 5 to 10mins.
I understand that the client would failover to the next available listener (On the next retry connection) if the node it is currently listening to has failed. Is there any configuration i should make to increase the failover efficiency?
Thanks for any advice.

Hi,
Server side failover is arranged by setting the remote_listener parameter.
Client side failover is set by using T(ransparent) A(pplication) F(ailover) (9i and higher)
or F(ast)C(onnection)F(ailover). Both are documented in the Net administrators manual for the version you didn't care to mention.
As far as I know, both TAF and FCF are not supported by the JDBC thin driver.
Sybrand Bakker
Senior Oracle DBA

RAC -process failover at node crash

Hi,
how to prevent running process(transaction) from termination in RAC while a node crashes ..
Ex: if there is a process running on node 1 and if it suddenly crashes in RAC how does we make node2 or node3 to pick it up and process or start the transaction again??
Thanks,

Hello,
Look at your tnsnames.ora entry and see if it configured to benefit from FAILOVER, you can aslo explore other available options options
http://stanford.edu/dept/itss/docs/oracle/10g/network.101/b10776/tnsnames.htm
myservice=
(DESCRIPTION=
   (SOURCE_ROUTE=yes)
   (ADDRESS=(PROTOCOL=tcp)(HOST=host1)(PORT=1630))    # <-- hop 1
   (ADDRESS_LIST=
     (FAILOVER=on)
     (LOAD_BALANCE=off)                                # <--- hop 2
     (ADDRESS=(PROTOCOL=tcp)(HOST=host2a)(PORT=1532))
     (ADDRESS=(PROTOCOL=tcp)(HOST=host2b)(PORT=1521)))
   (ADDRESS=(PROTOCOL=tcp)(HOST=host3)(PORT=1521))    # <-- hop 3
   (CONNECT_DATA=(SERVICE_NAME=myservice)))Another Example
MYSERVICE =
(DESCRIPTION =
(ADDRESS_LIST=
     (FAILOVER = on)
     (LOAD_BALANCE = on)
     (ADDRESS= (PROTOCOL = TCP)(HOST = server1)(PORT = 1521))
     (ADDRESS= (PROTOCOL = TCP)(HOST = server2)(PORT = 1521))
     (ADDRESS= (PROTOCOL = TCP)(HOST = server3)(PORT = 1521))
(CONNECT_DATA=
     (SERVICE_NAME =MYSERVICE)
     (FAILOVER_MODE =
         (BACKUP=server2)
         (TYPE=select)
         (METHOD=preconnect)
         (RETRIES=20)
         (DELAY=3)
) Regards
OrionNet
Edited by: OrionNet on Dec 18, 2008 3:12 PM

Apache failover service does not start - must zone name be the same?

Hello,
When I brought online apache data service, it starts successfully, then immediately fails. So the service starts up on the other node, then fails. No clear error message.
I'm thinking of potential causes:
- Could this be b/c the local zone names are different on both nodes? However, the zone paths are the same.
- The ftp-temp hostname is defined only on the 2 nodes /etc/hosts, not in DNS. In nsswitch.conf, I have "hosts: cluster files dns".
- I have the Bin_dir property set to /opt/xampp/bin. I can start up /opt/xampp/bin/apachectl start fine. Since I'm using xampp's apache, should I NOT be using SUNW.apache, but a generic data service?
- Does the Apache service must be under SMF control? Does sun cluster stop/start date services using SMF?
I'm running Solaris 10, SC 3.2
I want to fail over the Apache service betwn the local zones (not fail over the whole zone).
Here's what I did in setting up the resource from the global zone:
jupiter # clresource create -g ftp4-acme-com-rg -t SUNW.apache -p Bin_dir=/opt/xampp/bin -p
Network_resources_used=ftptemp -p Port_list=80/tcp,443/tcp,20/tcp,21/tcp xampp-apache
The ftp4-acme-com-rg was in unmanaged state.
jupiter # clresourcegroup online -emM ftp4-acme-com-rg
In /var/adm/messages:
Nov 24 13:25:47 jupiter Cluster.RGM.rgmd: [ID 443746 daemon.notice] resource ftptemp state on node jupiter:ftp-acme-zone2 change to R_ONLINE
Nov 24 13:25:49 jupiter Cluster.RGM.rgmd: [ID 784560 daemon.notice] resource xampp-apache status on node jupiter:ftp-acme-zone2 change to R_FM_ONLINE
Nov 24 13:25:49 jupiter Cluster.RGM.rgmd: [ID 922363 daemon.notice] resource xampp-apache status msg on node jupiter:ftp-acme-zone2 change to <Completed successfully.>
At this point, I saw the virtual ip plumbed up. Also, noticed 'online' states.
Resource Groups
Group Name Node Name State Suspended
Group: ftp4-acme-com-rg
jupiter:ftp-acme-zone2 Pending online No
Group: ftp4-acme-com-rg
pluto:ftp_zone Offline No
Resources --
Resource Name Node Name State Status Message
Resource: ftptemp jupiter:ftp-acme-zone2 Online Online - LogicalHostname online.
Resource: ftptemp pluto:ftp_zone Offline Offline
Resource: xampp-apache jupiter:ftp-acme-zone2 Starting Online - Completed successfully.
Resource: xampp-apache pluto:ftp_zone Offline Offline
The next message occurs 5 minutes later:
Nov 24 13:30:52 jupiter Cluster.RGM.rgmd: [ID 764140 daemon.error] Method <apache_svc_start> on resource <xampp-apache>, resource group <ftp4-acme-com-rg>, node <jupiter:ftp-acme-zone2>: Timeout.
Nov 24 13:30:52 jupiter Cluster.RGM.rgmd: [ID 443746 daemon.notice] resource xampp-apache state on node jupiter:ftp-acme-zone2 change to R_START_FAILED
Nov 24 13:30:52 jupiter Cluster.RGM.rgmd: [ID 529407 daemon.notice] resource group ftp4-acme-com-rg state on node jupiter:ftp-acme-zone2 change to RG_PENDING_OFF_START_FAILED
Same messages on the 2nd node.
This is the final status:
-- Resources --
Resource Name Node Name State Status Message
Resource: ftptemp jupiter:*ftp-acme-zone2* Offline Offline - LogicalHostname offline.
Resource: ftptemp pluto:*ftp_zone* Offline Offline - LogicalHostname offline.
Resource: xampp-apache jupiter:*ftp-acme-zone2* Offline Offline - Successfully stopped Apache Web Server.
Resource: xampp-apache pluto:*ftp_zone* Offline Offline - Successfully stopped Apache Web Server.
Any input would be appreciated.
Edited by: onetree on Nov 24, 2008 9:38 PM
Edited by: onetree on Nov 24, 2008 10:23 PM

Hi,
Did you eventually find the root cause/solution to this problem? I got into the same thing on my setup where I created 2 zones on a solaris 10 node. The apache on any of the 2 zones just dont seem to start up. In the end clrg failed after a few attempts on both zones. The time out is about 5 min after apache_svc_start was launched. I grepped httpd and no such process was launched. It does not like it is monitor issue. Logicalhost resource is started ok. so I am not sure if there is any security or access issue involed for apache_svc_start.
Mar 6 16:03:05 miwok.nms.fnc.fujitsu.com Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method
<apache_svc_start> for resource <apache-res>, resource group <apache-rg>, node <miwok.nms.fnc.fujitsu.com:ho
st1>, timeout <300> seconds
Mar 6 16:03:05 miwok.nms.fnc.fujitsu.com Cluster.RGM.global.rgmd: [ID 784560 daemon.notice] resource apache-
res status on node miwok.nms.fnc.fujitsu.com:host1 change to R_FM_UNKNOWN
Mar 6 16:03:05 miwok.nms.fnc.fujitsu.com Cluster.RGM.global.rgmd: [ID 922363 daemon.notice] resource apache-
res status msg on node miwok.nms.fnc.fujitsu.com:host1 change to <Starting>
Mar 6 16:08:07 miwok.nms.fnc.fujitsu.com Cluster.RGM.global.rgmd: [ID 764140 daemon.error] Method <apache_sv
c_start> on resource <apache-res>, resource group <apache-rg>, node <miwok.nms.fnc.fujitsu.com:host1>: Timeou
t.
One qestion is: do I have to install 3 separate copies of apache on global/zone1/zone2? I just had the default copy under /usr/apache.
Any info would be appreciated.
Mike

Re: (forte-users) loadbalancing a service object

Hi,
If you need to share the values between your So, you can find a sample on
http://perso.club-internet.fr/dnguyen
It was designed to show how to manage shared states between replicates using
fail over. It could also be used for load-bancing. A better way should be to
manage the names of service objects within Forte (with the application,
project name and compatibility level), but through the NameServer.
Hope this helps,
Daniel Nguyen
Freelance Forte Consultant
http://perso.club-internet.fr/dnguyen/
Xiao, Menghua a écrit:
Hello,
I tried to loadbalance a service object that has both methods and
attributes. The first copy of the service object works fine. However, all
the attributes of the second copy of the service object are nil when a
client try to use this copy. How can I make those two copies exactly the
same?
I am new to Forte. Any input will be greatly appreciated.
Menghua Xiao
Regional Transportation District,
Denver, Colorado, USA
For the archives, go to: http://lists.xpedior.com/forte-users and use
the login: forte and the password: archive. To unsubscribe, send in a new
email the word: 'Unsubscribe' to: forte-users-requestlists.xpedior.com

Hi,
If you need to share the values between your So, you can find a sample on
http://perso.club-internet.fr/dnguyen
It was designed to show how to manage shared states between replicates using
fail over. It could also be used for load-bancing. A better way should be to
manage the names of service objects within Forte (with the application,
project name and compatibility level), but through the NameServer.
Hope this helps,
Daniel Nguyen
Freelance Forte Consultant
http://perso.club-internet.fr/dnguyen/
Xiao, Menghua a écrit:
Hello,
I tried to loadbalance a service object that has both methods and
attributes. The first copy of the service object works fine. However, all
the attributes of the second copy of the service object are nil when a
client try to use this copy. How can I make those two copies exactly the
same?
I am new to Forte. Any input will be greatly appreciated.
Menghua Xiao
Regional Transportation District,
Denver, Colorado, USA
For the archives, go to: http://lists.xpedior.com/forte-users and use
the login: forte and the password: archive. To unsubscribe, send in a new
email the word: 'Unsubscribe' to: forte-users-requestlists.xpedior.com

How to connect to RAC with failover enabled?

I want to use instant client to connect to a RAC database. How can I accomplish that without tnsnames.ora file?

Not sure about RAC specifically. But you can always connect to an Oracle instance/listener without tnsname.ora, simply use the connection string which is the part to the right of "=" in each tnsname.ora entry.

Loadbalancing a service object

Hello,
I tried to loadbalance a service object that has both methods and
attributes. The first copy of the service object works fine. However, all
the attributes of the second copy of the service object are nil when a
client try to use this copy. How can I make those two copies exactly the
same?
I am new to Forte. Any input will be greatly appreciated.
Menghua Xiao
Regional Transportation District,
Denver, Colorado, USA

Menghua Xiao wrote:
I tried to loadbalance a service object that has both methods and
attributes. The first copy of the service object works fine. However, all
the attributes of the second copy of the service object are nil when a
client try to use this copy. How can I make those two copies exactly the
same?Ummmmm ...
How are you initializing these attributes? In your Init method? Do you need
to refer to any other SOs to do it? Forte does not guarantee the order in
which service objects are initialized. Referring to other SOs in the Init
method may not work, or it may work well for a long time and then suddenly
break when Forte decides to initialize things in a different order in
response to an upgrade, or to some random change in the application, or for
no known reason. If you're replicating, things could conceivably get even
more complex, with different replicates being initialized at different
times. And if one if the Init methods fails, Forte can't create the service
object, and your entire application fails to start.
Furthermore, Forte will tell you that attributes on a service object are a
bad idea. It's possible that they tell you this because there are bugs
around it that they're ashamed to tell you about (have you checked Forte's
web page?). Even if there are no bugs, there are a BUNCH of traps you can
fall into if you do this. The only time I consider it safe to have
attributes on a replicated SO is if those attributes can be reliably
maintained on that SO. This restricts it to:
* Database connection objects, and other such things that don't need to be
in synch.
* Static data that NEVER, NEVER changes.
The reason for these severe restrictions are that there are no reliable
ways to keep the two SOs in synch.
All that said, how can you get away with it? My religion says the best way
is to turn your attributes into virtual attributes. The quick and dirty way
to do this is:
1. Rename the attribute from AttrName to PrivateAttrName, and MAKE IT
PRIVATE.
2. Write a GetAttrName method that returns the attribute. This method
should check to see if the attribute is NIL, and if so load it. It ends
with "Return PrivateAttrName;".
3. Re-create AttrName as a virtual attribute, with "GetAttrName ()" (no
semicolon) as its get expression.
This approach gets the initialization of the attribute out of the Init
method, and instead forces initialization on the first reference. Hopefully
this solves all the problems. If it does not solve them, you ought at least
to get a useful error message out of your service object this way. I hope.
Good luck,
Tom Wyant

Rac loadbalancing/failover - services

Similar Messages

Maybe you are looking for