Reconciliation Common Errors and Resolution Tips Reconciliation process is the key to ensure data provider(s) generate the complete and correct production copy of CI and its relationship. When a CI is not identified or merged as expected from the corresponding RE job definitions, questions are raised around how we can find out what goes wrong. In order to trouble shoot the RE problems, you will need to have RE log for analysis. So the first step is to turn on Reconciliation Engine log: 1) Go to Reconciliation Engine tab in CMDB console. 2) From the left navigation pane, click “Settings”. 3) From the Server Settings dialogue window. Set “Debug Mode” to True and save the change. RE job logs are located at what’s defined in the “Log File Path” field on the Server Settings form. There’re many RE common problems which we list here along with error messages in the RE logs and their resolution tips. Found multiple matches in find-in-dataset during identification process. Symptom RE log message CI remains unidentified in the source dataset after identification process is finished. Sample error messages in the RE log: [Fri Feb 23 2007 13:40:05.7650] [ ERROR ] [TID: 002188] : Found multiple matches in dataset <BMC Asset> [Fri Feb 23 2007 13:40:05.7850] [ ERROR ] [TID: 002188] : Cannot identify the instance of class <BMC_ComputerSystem> with instance id <AG005056C00008wQfeRQ66iBAQ0owA> Root Cause Resolution Find-in-dataset (BMC.Asset in this example) has two or more CIs which share the same value on identification attribute. 1) Search for CI based on instance id (In our sample log, instance id is AG005056C00008wQfeRQ66iBAQ0owA.). 2) Search CIs from find-in-dataset (BMC.Asset) based on identification attribute values of the CI from step 1). For example, if TokenId attribute is used in the identification rule, search for CI based on dataset id equals find-in-dataset id and TokenId equals to what you see from the CI instance from step 1). 3) If multiple CI instances are found from step 2), delete the duplicates entry so only one CI exists in find-in-dataset; if multiple CI instances refer to different CIs, re-examine the identification rule or the data quality and make appropriate adjustment. 4) Run identification process to get the CI identified. Unique index violation during identification process. Symptom RE log message CI remains unidentified in the source dataset after identification process is finished. Sample error messages in the RE log: [Fri Feb 23 2007 14:17:37.1120] [ ERROR ] [TID: 002188] : ARERR[120092] The dataset ID and Reconciliation Identity combination is not unique. Root Cause Resolution There’s a duplicate identified CI in the source dataset. 1) Search for CI based on instance id. 2) Search CIs from find-in-dataset (BMC.Asset) based on identification attribute values of the CI from step 1). For example, if TokenId attribute is used in the identification rule, search for CI based on dataset id equals find-in-dataset id and TokenId equals to what you see from the CI instance from step 1). Make a copy of ReconciliatonIdentity attribute value if a CI is found. 3) Search for CI based on source dataset id and ReconciliationIdentity value from step 2). 4) Check to see whether the found CI is marked as softdeleted. If it’s marked as soft-deleted, delete the record; if it’s not marked as soft-deleted, re-examine the identification rule and make appropriate modifications. 5) Run identification process to get the CI identified. Missing end points during merge process. Symptom Relationship CI can’t be found from BMC.Asset after merge process is finished while such identified CI exists in source dataset. RE log message [Thu Jan 25 2007 11:58:27.9156] [ ERROR ] [TID: 000008] : Cannot find the endpoint of relationship::BMC_Application [Thu Jan 25 2007 11:58:27.9161] [ ERROR ] [TID: 000008] : ARERR[0] Cannot find the endpoint of relationship::BMC_Application [Thu Jan 25 2007 11:58:27.9162] [ ERROR ] [TID: 000008] : Merging of record failed: Root Cause Modification to the CI from source dataset without modifying corresponding relationship CIs. For example, resetting reconciliation identity of BMC_ComputerSystem CI without resetting the associated relationship CIs’ reconciliation identity. 1) Reset reconciliation identity to be “0” for all instances in the specific dataset. 2) Re-run identification and merge process. Resolution Incremental merge becomes full merge. Symptom Only a few CIs were modified since last RE merge. But the ongoing merge takes much longer than usual. RE log message [Fri Feb 23 2007 13:40:08.6690] [ INFO ] [TID: 003172] : Merging class: BMC_ComputerSystem [Fri Feb 23 2007 13:40:08.7190] [ INFO ] [TID: 003172] : Total number of records found which are to be merged are <527867> This number is much bigger than anticipated. Root Cause Merge activity’s “Last Activity Time” is reset to NULL for the following reasons: 1) The previous RE job run is aborted. 2) RE job definition such as schedule or merge algorithm is changed. Resolution: 1) Open “RE:Activity” form with Remedy Administrator Tool. Scroll down to the bottom part of the form to view page holder with pages System_tab & Debug_tab. Make sure the page holder is not set to “Hidden”. 2) With the Remedy User Tool, open “Reconciliation Activities” search form and search for the merge activity and get the activity id from the “instanceId” field. 3) With Remedy User Tool, open search from “RE:Activity_Runs” and fill in “Activity Instance ID” field with activity id retrieved from previous step and search. 4) Fill in the appropriate value in both “Last Activity Start Time” and “Last Activity End Time” fields. 5) Now RE merge process should perform on delta basis. Merge algorithm selection. Symptom The same CI seems to be merged more than once during merge process. RE log message [Fri Feb 23 2007 13:40:08.6090] [ INFO ] [TID: 003172] : Algorithm: Including child CIs and committing together Root Cause Resolution Unnecessary complicated merge algorithm selection. If no selectivity is needed for merge activity, choose “Stand Alone” algorithm. Run RE with a single thread. Symptom RE has poor performance in all activities. RE log message Root Cause Resolution RE log only shows two unique TIDs (Thread ID). You forget that RE can be set with multiple threads. Use Remedy Admin Tool to set RE to be multi-threaded. RE has poor performance due to incorrect database settings. Symptom RE log message Root Cause Resolution RE has poor performance in all activities. Oracle 10g has default cursor_sharing set to be ‘exact’. Oracle database instance has small memory allocation. SQL server has small tempdb space allocation. Set cursor_sharing to be ‘force’ or ‘similar’ to boost performance by up to 30%. 2) Get DBA involved ( run statspack to identify the database performance bottleneck ). 3) Allocate more dedicated fast disk space for SQL Server tempdb. 1) 2) 3) 1) Database indexing can be critical to RE performance. These areas require good attention. Identification attributes are indexed properly. The out of box BMC Atrium indexes are not modified or deleted. Index statistics are updated regularly. RE aborted due to temporary disk space limit. Symptom RE log message RE aborted. Root Cause Reconciliation process leads AR API requesting temporary disk space which is not available in the file system. Add more disk space to the server TEMP disk or redirect TEMP to another disk which has more space. Resolution [Wed Nov 01 02:47:50 2006] [ ERROR ] [TID: 002836] : ARERR[316] Failure with a temporary file [Wed Nov 01 02:47:50 2006] [ ERROR ] [TID: 002836] : No space left on device No memory available during reconciliation process. Symptom RE log message Root Cause Resolution RE aborted. ARERR 300 Malloc failed on server Memory usage is too high to allow new memory allocation. 1) Make sure the server has enough memory resource. We recommend CMDB/RE is configured to be on a separate server so there will be no share resource for CMDB/RE. 2) Check virtual memory setting to make sure there’s enough virtual memory space. Due to intensive data manipulation, there may be needs to use more than what’s available in system physical memory. 3) Fine tune RE multi-thread number. When using “include child CIs” algorithms, reducing the number of RE threads will decrease the required memory by Reconciliation process. AR server timeout issue. Symptom RE log message RE aborted or RE seems to be hanging. [Mon Oct 16 20:10:43 2006] [ ERROR ] [TID: 005256] : Failed in getting instances from class <BMC_PRODUCT>:: [Mon Oct 16 20:10:43 2006] [ ERROR ] [TID: 005256] : ARERR[94] Timeout during database query -- consider using more specific search criteria to narrow the results, and retry the operation [Thu Nov 9 10:51:23 2006] [ ERROR ] [TID: 000008] : Error while loading instances with api AROSGetMultipleInstances():: [Thu Nov 9 10:51:23 2006] [ ERROR ] [TID: 000008] : ARERR[93] Timeout during data retrieval due to busy server -- retry the operation Root Cause Resolution Bad SQL statement by AR or RE. 1) There’s a known issue with AR generated SQL statement for AR joint form if there’re attachment/attachment pool fields involved. Workaround can be either to delete the unused attachment/attachment pool fields or to modify the joint form SQL statement manually. 2) If timeout happens when RE is merging millions of CIs from a specific class, we can divide the merge process into smaller RE jobs to limit the data retrieved from AR server. For example, we can define multiple jobs of merging BMC_Product class grouping by alphabetical name ranges and having one job trigger the other one to start with “execute job” activity sequentially. 3) Consult with DBA for troubleshooting.
© Copyright 2026 Paperzz