KB-1419 Negative KDB numbering

Symptom

KDB file numbering for one or more engines moves into the negatives. This may manifest in:

• the names of actual .kdb files on disk
• the checkpointing messages in the corresponding gw_*.log
• the DB ID referenced in the corresponding db_*.log

For example, if the process execution01 engine is the one in question, you may notice the following:

• A negative .kdb file in <APPIAN_HOME>/server/process/exec/01/gw1/: pe-2147483646.kdb.
• A reference in the gw_PX001*.log: [PX001] INFO .a.gw.ssa "Server switchover to database ID -2147483646".
• A reference in the db_PX001*.log: [PX001] {pe-2147483646.kdb ,4405}.

Cause

This issue can occur after improperly resetting a particular engine's KDBs or if extraneous KDB files are manually introduced.

Action

1. Shut down the engine with the negative KDB numbering.
2. Rename any negative KDBs to positive numbers, ensuring that the most recent KDB has the highest number.
3. Check that all KDB file names look normal - i.e. no remaining negative numbers or manually introduced KDBs such as pe.kdb.
4. Start the engine and confirm that KDB numbering is positive once again.

Warning: Renaming .kdb files is not supported beyond what is documented in Appian documentation, the Appian Playbook, and KB articles.

Affected Versions

This article applies to all self-managed versions of Appian.

Last Reviewed: November 2017