Article Content
Article Number | 000016564 |
Applies To | Re-run initial_multinode_install.cmd will not fix the issue. |
Issue | This issue would happen in EA structure, observed on 4.1 patch 8. The CA node might have a incorrect mobilink subscription name of LC instead of CA node name. <systemname>_sync_succeeded file will have incorrect node name,(ex site-LC1 instead of site-CA1) |
Cause | It is a bug that introduced in ECE-15 NICDB replication improvements. Those commands are internally using the 'node' value from the 'dbsettings' table as the system name for replication name, it should be using the 'systemName'. |
Resolution | Jira ticket ECE-2433 is opened for this issue. The work around fix for customers with the issue: ? [CA] Stop all services on the CA node except for the NIC DB service. (the replication client should be off) o You should make this the offline collector if possible o This also assumes that the CA is currently sync?d with the DS node using the wrong subscription name. ? [CA] Run database\cmd\remove_mobilink_subscription.cmd <sitename>-LC1 ? [DS] Run database\cmd\remote_sysnc_user.cmd <sitename>-LC1 o This will clean up the ml_subscription, ml_user, and ml_sync_user table entries assocated with <sitename>-LC1 on the DS ? [CA] Update the db with the ECE2433WorkAroundFix.sql (can be found in ECE-2433) o Installing this in the CA by copying to the database\sql\patches directory and then executing database\cmd\install_patches.cmd. ? [CA] Run database\cmd\create_mobilink_subscription <sitename>-DS1 ? [CA] Start Replication Service ? [CA] Verify replication ok, and using correct name. ? [CA] Start remaining services if any. Verify _<systemname>-sync_succeeded file has the correct name for the CA node. |
Legacy Article ID | a60972 |