RMAN-06004: ORACLE error from recovery catalog database: ORA-03135: connection lost contact

RMAN-06004: ORACLE error from recovery catalog database: ORA-03135: connection lost contact

If there is a firewall between the catalog database and target  database then it may be possible that your backup is getting errors below


Starting backup at 01-MAY-10
current log archived
released channel: ch1
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of backup plus archivelog command at 05/01/2010 15:29:59
RMAN-03014: implicit resync of recovery catalog failed
RMAN-06004: ORACLE error from recovery catalog database: ORA-03135: connection lost contact
ORACLE error from recovery catalog database: ORA-03114: not connected to ORACLE

*** what is happening here is that your catalog database machine is dropping the connection of RMAN after
it reaches tcp_keepalive_intvl or timeout value set in firewall. So the RMAN session is dropped by the machine and
backup throws errors above.

Here is what can be done  to overcome the problem

Set  SQLNET.EXPIRE_TIME in your sqlnet.ora of catalog database.You can set it to time below your
tcp timeout setting of the firewall and the machine itself.

Let's say RMAN takes two hours to finish database backup and tcp_keepalive_intv is set to 75 on your catalog database machine then following settings can be user

for example

SQLNET.EXPIRE_TIME=20

Second , set your tcp_keepalive_intvl to higher number on your database server machine

Here is an example for Linux Machine


cat /proc/sys/net/ipv4/tcp_keepalive_intvl

default is 75

to change it to  50 [ this value depends on the time RMAN takes to backup your database]

echo 50 > /proc/sys/net/ipv4/tcp_keepalive_intvl


to verify the value change

cat /proc/sys/net/ipv4/tcp_keepalive_intvl

You can set above value in /etc/sysctl.conf also