Part 10 (Full Master Edition): Repository (SRF), File System & Specialty Engines
Storage & Metadata Architecture Diagnostics: Comprehensive Scenario Analysis covering Compiled Repository Schema Discrepancies (SBL-REP-00100), Workspace Merge Conflicts & Rebase Collisions (SBL-REP-00142), File System Storage Directory Mount Drops (SBL-FS-00101), Capacity Limits & Latency Timeouts (SBL-FS-00115), and Loyalty Engine Transaction Exceptions (SBL-LOY-00105).
SBL-REP-00100 error occurs when an Application Object Manager task attempts to query an object (Applet, Business Component, View) whose compiled definition in the active target repository (`siebel.srf` or delivered Workspace metadata) diverges from the underlying database physical table structure (`S_COLUMN`, `S_TABLE`). This leads to invalid SQL query generation or runtime C++ pointer exceptions inside `siebmtshmw`.
A developer delivered a schema workspace modifying table columns, but the physical database schema was not updated, causing Object Managers to crash when querying non-existent columns.
- In Web Tools, verify that the active workspace has been successfully delivered to `MAIN`.
- Run Database Schema Sync via `dbdict` / `ddlctl` utility or perform **Apply/DDL** on target tables from Siebel Tools/Web Tools.
App server processes serve stale repository definitions from local process memory after a deployment, throwing dynamic SQL runtime exceptions.
- Clear the server-side dictionary cache and bounce affected Object Managers using `srvrmgr`.
One node in an Enterprise cluster runs an older `siebel.srf` binary file while other nodes run a newly deployed version, causing intermittent load-balancer thread crashes.
- Ensure deployment scripts sync the compiled `siebel.srf` file identically across all server nodes in `$SIEBEL_ROOT/siebsvr/objects/
`.
A field marked as `Nullable = FALSE` in Siebel repository allows `NULL` values in the physical database definition, causing runtime C++ integrity check failures.
- Regenerate physical database schema definitions or update mandatory column attributes in Siebel Tools and execute `Apply DDL`.
SBL-REP-00142 merge conflict is thrown in Web Tools when a developer attempts to deliver a child workspace to `MAIN`, but a concurrent delivery modified the same repository object attributes, requiring a manual workspace rebase.
Attempting to deliver a child workspace to `MAIN` fails because another developer delivered changes to the same Business Component or Applet properties in the interim.
- In Web Tools, open the target workspace and click **Rebase**.
- Review conflicting object properties in the **Merge Conflicts** panel and select target attribute resolution (`Mine` vs `Latest`).
- Re-test the workspace locally prior to re-initiating delivery to `MAIN`.
An application server process or developer session crashed during a workspace delivery, leaving the repository branch permanently locked in `DELIVERY_IN_PROGRESS`.
- Run SQL diagnostics to inspect workspace tree locks in `S_WORKSPACE` and reset stuck delivery flags.
A developer workspace created several release cycles ago attempts to rebase against a substantially modified parent branch, causing automated merge engine failure.
- Export developer workspace object changes to XML via Web Tools.
- Discard the stale workspace, create a new child workspace from latest `MAIN`, and re-import objects.
Delivering a workspace fails because an applet control or business component field references a join or picklist object defined in a separate un-delivered workspace.
- Sequence workspace deliveries so underlying data structures and picklists are delivered to `MAIN` prior to dependent UI component workspaces.
SBL-FS-00101 error is raised by the `File System Manager` (`FSManager`) component when it fails to write, open, or decompress an attachment file (`.SAF`) on the Siebel File System share (`$SIEBEL_FS`). Triggers include unmounted NFS/SMB network shares, invalid file path parameter configurations (`FileSystem`), stale OS file locks, or missing write permissions for the OS application service account.
The File System Manager (`FSManager`) fails to retrieve or store attachment files (`.SAF`) because the network storage mount point unmounted or timed out.
- Verify the status of the file system share on the Linux/Windows host running the Application Server.
- Ensure the `siebel` OS service account possesses full read/write permissions across all subdirectories.
The Enterprise parameter `FileSystem` points to an invalid UNC network path or non-existent local directory structure.
- Query and update the `FileSystem` parameter across Enterprise components using `srvrmgr`.
Uploading or downloading compressed attachment files (`.SAF`) fails because underlying OS binary compression libraries are unreadable by the `siebel` service account.
- Check read and execute permissions on `$SIEBEL_ROOT/siebsvr/bin/minizip` or OS compression binaries. Grant execute rights to the server service account.
A record in `S_ACCNT_ATT` or `S_PARTY_ATT` references a file name token that was deleted or renamed directly on the storage file system.
- Run diagnostic SQL to cross-reference unlinked database attachment tokens against physical storage.
SBL-FS-00115 error occurs when the primary File System volume reaches 100% capacity or disk latency on the underlying SAN/NAS causes `FileSystemManager` network connection calls to exceed the `FSConnectTimeout` threshold.
Attachment write requests fail across all application components because the primary NAS/SAN disk volume ran out of free block space.
- Configure multi-directory storage paths separated by semicolons in the `FileSystem` parameter to distribute attachments across multiple storage pools.
Heavy concurrent attachment read/write operations cause file transfer latency to exceed the default `FSConnectTimeout` threshold, dropping client sessions.
- Increase the `FSConnectTimeout` parameter in `srvrmgr`.
Deleted business records leave unlinked `.SAF` attachment files behind on the file system, consuming gigabytes of disk space over time.
- Run orphaned file system purge scripts (`sfsclean`) to cross-reference database attachment tables against physical `.SAF` files and free disk space.
Storing over 65,000 files in a single attachment subdirectory (e.g., `/att`) degrades directory lookup performance and triggers I/O execution errors.
- Enable nested directory storage parameters in Siebel File System configuration to balance files into subfolders (e.g., `/att/01`, `/att/02`).
SBL-LOY-00105 error occurs when the **Loyalty Engine** component (`LoyEnginePrc`) encounters an unhandled rule processing error while evaluating transaction rows (`S_LOY_TXN`). Causes include circular tier assessment logic, missing point type bucket mappings, corrupt member transaction histories (`S_LOY_MEMBER`), or unindexed candidate promotion queries.
Transaction records remain stuck in `Processing` or `Error` state due to unhandled tier evaluation or point calculation logic exceptions inside `LoyEnginePrc`.
- Run SQL diagnostics to locate stuck or errored Loyalty transaction records in `S_LOY_TXN`.
- Reset failed transaction status back to `Queued` for re-processing.
Peak transaction import volume overwhelms the single `LoyEnginePrc` task thread, causing engine queue backlogs and processing timeouts.
- Adjust thread allocations and task scaling parameters for `LoyEnginePrc` via `srvrmgr`.
A loyalty program promotion attempts to assign points to an unconfigured Point Type bucket or active period, halting transaction rule processing.
- Navigate to **Administration - Loyalty > Program Design**.
- Ensure all Point Types, Buckets, and Action Sets assigned to active promotions are enabled and valid.
A misconfigured tier assessment rule triggers infinite evaluation loops between tier upgrade and downgrade triggers, crashing `LoyEnginePrc` threads.
- Inspect loyalty engine log files for recursive rule execution, deactivate target promotion, and fix tier assessment criteria logic.