|Applies To||RSA ACE/Server 5.0.1 (no longer supported as of 8-15-2004)|
UNIX (AIX, HP-UX, Solaris)
|Issue||RSA ACE/Server replica will not start and getting error when running sdconnect|
Error: "BROKER 0: Invalid BI file, before-image file is zero length. (877)"
Error: "SYSTEM ERROR: Invalid shared memory segment identifier"
Error: "Save file named core for analysis by Progress Software Corporation. (439)"
|Cause||The database on the replica has become corrupt|
|Resolution||To correct this issue, create a new Replica Package and install manually on the Replica server not starting:|
On the Primary:
1. Run ACE/Server Database Administration
2. Select the System menu
3. Deselect the "Allow DBPUSH" option which is set on as default
On the Primary:
1. Delete all files and directories from under the $REP_ACE directory
2. Change the current directory to the $USR_ACE directory. Type:
3. When prompted, enter the name of the affected Replica server. Type:
4. When prompted, enter the name of the affected Replica server
On the Replica:
1. Copy the contents of the $REP_ACE data directory on the Primary to the $VAR_ACE directory on the Replica server
2. (UNIX only) Reset the permissions on the database files by running "sdsetup -config" on the Replica server
3. Start the ACE/Server on the Primary and Replica
If this happens on the Primary server, you will need to chose a recovery option:
1. Select the most up-to-date Replica server and carry out the Disaster Recovery procedure on the Primary server
2. Restore the database on the Primary server from a backup
3. If only the log database is corrupt and you have determined you do not need to keep the log records, sdnewdb the log database
NOTE: After restoring the Primary server, create and install a Replica Package on each of the Replica servers
If this problem occurs after creating a new Replica package, please contact RSA Security Technical Support.
|Legacy Article ID||a12948|