Views:


Symptoms



An Oracle Point in Time Restore will fail with the time out errors below within 2 hours of execution.

ERROR     Remote command(bash /tmp/cdm_guestapps_cdmagent/oracletools.sh /tmp/fad54caf-726b-4456-9bb3-9df35677fafc) execution failed with timeout)
ERROR     Remote command(bash /tmp/cdm_guestapps_cdmagent/oracletools.sh /tmp/fad54caf-726b-4456-9bb3-9df35677fafc) execution failed. Exception(java.io.IOException: Stream closed)
WARN       The command executed on the remote host failed with return code of -1
ERROR     Restore databases: Failed to restore Oracle databases (live) on server (oraqa1). Exception (com.syncsort.dp.xsb.commons.XSBException) reports: (Failed to retrieve output.json from remote host. e=2: No such file)
ERROR     Restore databases: Failed to restore Oracle databases (live).
 

Resolution



If this is a busy Oracle server, the point in time restore will need more time to complete the restore process. Increase the SSH Timeout setting on the Appliance to resolve this 2 hour time out limit.

On the appliance, edit the file /opt/virgo/repository/ecx-usr/com.catalogic.ecx.deploy.vmware.ecxvmdeployer.json. 
Edit or insert the "sshTimeout" value which is available in milliseconds, and set it to 86400000 (24 hours). 
Example:

"transferTimeout":3600000, 
"sshTimeout":86400000, 
"rubyPath":"/usr/bin/ruby" 

Reboot the appliance and re-run the restore job.  It will now continue to wait past the 2 hour time out limit.
Related Products: ECX