Owner's Manual

Deployment Guide 27
27 On one node , type:
srvctl status database -d dbname
where dbname is the global identifier name that you defined for the database in DBCA.
If the database instances are running, confirmation appears on the screen.
If the database instances are
not running, type:
srvctl start database -d dbname
where dbname is the global identifier name that you defined for the database in DBCA.
RAC Post Deployment Fixes and Patches
This section provides the required fixes and patch information for deploying 10 g RAC.
Reconfiguring the CSS Miscount for Proper EMC PowerPath Failover
When an HBA, switch, or EMC Storage Processor (SP) failure occurs, the total PowerPath failover
time to an alternate device may exceed 105 seconds. The default CSS disk time-out for 10
g R1
version 10.1.0.3 is 45 seconds. To ensure that the PowerPath failover procedure functions correctly,
increase the CSS time-out to 120 seconds.
To increase the CSS time-out:
1 Shutdown the database and CRS on all nodes except on one node.
2 On the running node, log in as user root and type:
/opt/oracle/product/10.1.0/crs_1/bin/crsctl set css misscount
120
3 Reboot all nodes for the CSS setting to take effect.
For more information, see Oracle Metalink Note 294430.1.
Configuring Cluster Ready Services
Intermittently while configuring Cluster Ready Services (CRS) on Oracle 10 g for EM64T on Red
Hat Enterprise Linux (version 3), the
CRS_STAT command hangs and does not return the CRS
daemon status. Oracle has identified this issue as Oracle BUG 4039296.
To resolve this issue, download the following patches from the Oracle Metalink site:
• Patch number 4039296 with Linux x86-64 (AMD)
• Patch number 2617419 with Linux x86-64 as the platform
Using the Oracle Opatch utility, install the downloaded patches on all of the cluster nodes.