SUMMARY
Recovery Series appliances may experience backup failures caused by space issues.
ISSUE
Recovery Series appliances may experience backup failures caused by space issues related to inappropriate assignment of metadata backups to added external storage devices.
RESOLUTION
To confirm if you are effected by this issue, view the file /usr/bp/bpinit/master.ini after logging into your appliance with PuTTY.
head -10 /usr/bp/bpinit/master.iniIn the [server Information] section the first line is "server1"
If the name on this line is not equal to your appliances network hostname but is instead the name of an external storage device, then you have encountered this issue.
This issue is also known to only occur if ETH0 has been included in a network bond against supported practice.
To resolve this issue, upgrade your appliance to version 10.1.
CAUSE
An issue with the update_dpu_script may cause system metadata backups to associate themselves with added storage rather than the appliance itself.
Per our bonding KB ETH0 is never permitted to be in a bond, and appliances that include 2 or fewer network adapters are not supported for bonding without the purchase of an additional 4 port adapter. UBs should never have virtual adapters bonded as bonding would be done at the hypervisor level, not the guest level. Thus this scenario had never been tested or certified resulting in a defect with unsupported configurations. Code has been modified to prevent the metadata impact and reflect proper master.ini information, however if you have included ETH0 in a bond this must be removed and appropriate adapters configured in your bond. Making changes to adapter use of ETH0 can have license activation impacts, Unitrends support will require SSH access to your appliance remotely through a support tunnel to resolve improper bonding configurations and license issues.