IMPAX 6.5.1 Server Knowledge Base home > Troubleshooting: IMPAX Server > Troubleshooting: Oracle Server
Issue
After Oracle Data Guard has failed over to the standby server, there is a long delay before Solaris clients such as Network Gateways can connect to the standby database. This delay can be up to 3 minutes long. During this time, IMPAX essentially ceases to function.
Details
The delay is caused by the TCP/IP settings. You can significantly reduce this time interval by changing the TCP/IP values on the Oracle database’s Solaris clients.
Solution
![]() | Important! This solution applies to Solaris servers only. This procedure is not necessary on Windows clients in a mixed-host cluster. |
On each of the Oracle database's Solaris clients, change the TCP/IP values as follows:
Log in to one of the Oracle database's Solaris clients and open a command prompt.
View the current TCP settings by typing:
ndd -get /dev/tcp tcp_ip_abort_cinterval
ndd -get /dev/tcp tcp_ip_abort_interval
ndd -get /dev/tcp tcp_keepalive_interval
Record these values in case you have to reset them.
Change the current TCP settings by typing:
ndd -set /dev/tcp tcp_ip_abort_cinterval 10000
ndd -set /dev/tcp tcp_ip_abort_interval 60000
ndd -set /dev/tcp tcp_keepalive_interval 240000
After modifying the TCP/IP values, the client will connect to the standby Oracle database much faster than before.
Topic number: 111472 Applies to: IMPAX 6.5.1 Server Knowledge Base |