I have a jenkins setup running on solaris. I have configured a job which runs a .bat file, using the jenkins ssh plugin, on a remote Windows 2008 R2 server having an oracle database. The .bat files does an RMAN restore in the oracle Database. I have an openssh server running on Windows 2008 R2 server. When I do a restore on smaller database, which takes less than 30 mins to run, everything works fine. But when the databse is large and the restore takes over 3 hours, the jenkins job hangs even though the rman restore srated by the job completes successfully on the server.
This is the command I run in jenkins ssh plugin command window:
/cygdrive/C/OpenSSH/bin/SSH_ImportTEST_DB.bat TEST3
This is what the .bat file does.
- Prepare the DB for RMAN restore(runs few sql commands as system user)
- Run the RMAN restore
- Run few more sql commands to let the Database be available for all users, reset the passwords.
Step 1 and 2 work fine but the jenkins job just shows as running step 2 even though it has completed on remote server. It never executes step 3 if the rman restore runs over 3 hours.
X:\oracle\admin\TEST3\scripts>rman auxiliary cmdfile=X:\oracle\admin\TEST3\scripts\restore_db2.sql log=Y:\oracle\NightlyBackup\Prod\restore_MSIPTST3.log
RMAN> 2>
This is what the restore_db2.sql contains:
DUPLICATE DATABASE TO TEST3 BACKUP LOCATION 'Y:\oracle\NightlyBackup\Prod';
exit;
I have tried the restore_db2.sql without the exit command, but that doeson't help either.
I have configured ServerAliveInterval 60 in ssh_config on the server where jenkins is running. I have also configured ClientAliveInterval 60 in sshd_config on the openssh server to rule out timeouts.
Can someone assist?
Set up your windows client as jenkins slave because cygwin openssh is not a proper way for remote jobs. In your case I would recommend the Java Web Start option. An configuration example is published on: https://wiki.jenkins-ci.org/display/JENKINS/Step+by+step+guide+to+set+up+master+and+slave+machines